Part I.

One is the dystopian Terminator version where the machines become fully autonomous and take over, eliminating existential human threats. Machines are smarter and faster, programmed to be heuristic, deterministic and predictive. To break this down - and to set the context and scale of change in order understand the impact of the gaps in enforcing it - I asked Gemini to break it down.

It still lacks the edge of human judgment, human experience and is limited to the it was fed.

I will include the invariably better products and services this has led to, at the end of this post.

The second version is where governments, business leaders invest in retraining the workforce, proactively bringing in the new tech (without proven impact with their data and architecture, and littered unconnected use POCs) and keeping humans in the loop, while keeping the experimental loop open. I admit that this is utopia compared to reality and wishful thinking. Two highly successful American VC and founder at the mention of reskilling and retooling the American workforce, emphatically said: that’s not going to happen.

Why not? Are not humans the best competitive AI strategy in the arms race?

Yes, I fail to understand why we are determined to struggle, with pockets of AI talent density in a handful of zip codes, industries and institutions.

Instead, companies that laid off teams and scores of human talent are realising their error and hiring people back. The human judgment gap is real. Ford hired back engineers when AI failed to meet quality checks, because the cost of failure in trust and reputation is too high. Why were the engineers let go off in the first place, without testing? Employers are regretting the shotgun decision globally. The human in the loop, judgement calls, and institutional knowledge cannot be crunched by resource hungry machines.

Part II.

Now that we understand the context, impact on jobs, and leadership decision. Lets shift to the renewed respectability of change management as discpline. We will study it with two high growth startups Distyl AI’s has a highly successful “Outcome Ownership” approach and their partnership with Google Cloud and Writer AI’s enterprise data approach.

The traditional SaaS Proof of Concepts (POCs) and the legacy Change Management models are failing in the age of AI.

The AI Change Crisis: Why Prosci is Failing the Agentic Enterprise (and What Comes Next)

This post has been in the making for longer than I care to admit. With the recent rise in Change Management as AI adoption increases and internal resistance makes news, roles for Change Management leads have increased.

My Change Management career was in the Strategy & Change Management practice of BearingPoint Consulting. The complex large-scale projects implementing ERP, invested in change expertise, with client leadership understanding the risks of ignoring workforce impact.

The Prosci ADKAR model was upgraded to the Strategic Change Management wheel inspired by Kotter’s 8-Step Change Model. The skill lies in knowing which one to pull from and when.

An overview of best known models.

• People-Focused Models

  • Prosci ADKAR Model: Emphasizes that organizations change one person at a time. It breaks change into five stages: Awareness of the need for change, Desire to support it, Knowledge of how to change, Ability to implement new skills, and Reinforcement to sustain it.

  • William Bridges’ Transition Model focuses on the psychological and emotional journey of employees, differentiating between the event (change) and the internal psychological process (transition): the End, the Neutral Zone, and the New Beginning.

  • Kubler-Ross Change Curve, adapted from the stages of grief, helps leaders understand and manage employee emotional responses, typically running from shock and denial to bargaining, depression, and eventual acceptance.

• Process-Focused Models

  • Lewin’s Change Management Model: A classic, high-level, three-step framework that guides behaviors by Unfreezing (preparing for change), Changing (implementing new processes), and Refreezing (stabilizing and normalizing the change).

  • Kotter’s 8-Step Change Model: A structured, top-down framework designed to create urgency, build a coalition, establish a vision, and anchor new behaviors into the corporate culture.

• System & Holistic Models

  • McKinsey 7-S Framework: Focuses on aligning seven interconnected elements of an organization (Strategy, Structure, Systems, Shared values, Style, Staff, and Skills) to ensure the organization is well-structured to achieve its objectives.

These models are fixed in time and designed for a static event. Reactionary approach over an adaptive iterative one, with design and decision controls close to floor, are non-negotiables. In the visuals for the paid subscribers, see AI impact and notes on adaptation.

The AI Paradox

Over 94% of the C-suite uses AI daily, yet a staggering 79% of organizations report massive friction, with 54% of executives admitting AI adoption is “tearing their companies apart” (source: Writer AI Enterprise Survey).

The Problem: We are no longer dealing with simple writing assistants or passive software. We are deploying autonomous AI agents that radically rewrite workflow architectures.

The Core Thesis: Traditional Organizational Change Management (OCM) is facing an ideological schism. The linear tools we use to manage human adoption are utterly unequipped for the velocity of an AI-native operational shift. To survive, change management must merge with AI Design Thinking, treating AI not as a tool to train for, but as a digital teammate to build with.

The Two Schools of Thought: The Ideological Rift

The two factions currently battling for control over enterprise transformation lies between the traditional legacy mindset and the radical AI-natives.

School 1: The Legacy Traditionalists: The Prosci Purists & Traditional SaaS POCs

The mindset is “Change is change”. Therefore, rely on ADKAR (Awareness, Desire, Knowledge, Ability, Reinforcement).

• The Fatal Flaw: This group relies on long-tail training manuals and the classic SaaS ‘Proof of Concept’ (POC) checklist. But as Distyl AI’s CEO Arjun Prakash points out, the traditional SaaS POC model is a terrible fit for AI because requirements change the second the technology touches real-world enterprise data.

• The Incentive Problem: Traditional practitioners have zero organizational incentive to learn these tools. Their billable hours and legacy certifications rely on slow, highly manual, long-tail human processes, not outcomes.

• The Result: Performative Change, where 75% of executives admit their AI strategy is more for show than actual guidance, and employees pretend to adopt tools to pass corporate audits without changing how business value is actually generated.

I have witnessed this firsthand where a strategy team responsible for transformation as a result of AI tools, did not themselves understand nor use the tools. A training course is not the solution.

School 2: The Radical AI-Natives: The Operational Redesigners

The Mindset: “AI is not software you learn; it is an active teammate that alters decision rights and role definitions.”

• The Strategy: Drawing from frontier trend-setters like Writer AI and Distyl AI, this school focuses on structural transformation and rapid outcome delivery (weeks, not quarters) rather than incremental productivity gains.

• The Shift: They move from process optimization (making an old step faster) to process redesign (eliminating the step entirely via autonomous AI workflows).

TL;DR: They are helping incumbents go AI-native by rebuilding the processes ground up, rather than introducing bandaid plastered solutions that begin to emulate Frankenstein, even with orchestration.

Enterprise Proof Points: Success Stories from the Frontier

Validation on why School 1 is losing ground by showcasing real-world, high-stakes deployments driven by frontier startups.

Case Study A: Distyl AI & Google Cloud GTM Partnership (The Rearchitects)

• The Strategy: Distyl AI’s strategic Go-To-Market partnership with Google Cloud will accelerate enterprise AI transformation. Distyl bypasses traditional consulting by using Forward-Deployed Engineers (FDEs) who co-design systems directly with business leaders and share accountability via outcome-based pricing. Source: Distyl.

• Real-World Impact

  • Manufacturing: A Fortune 50 hardware manufacturer went from manual tracking to root-causing 1,500+ supply chain disruptions daily, cutting resolution time by up to 47%.

  • Finance/Insurance: An auto finance lender achieved a 93% cost reduction for loan origination by moving from kickoff to agentic fraud detection in just one week.

  • Telecom: A Fortune 100 operator saw 75%+ of customer interactions handled via AI, projecting over $ 200M in Opex savings.

Case Study B: Writer AI: The Democratizers of Workflows

• The Strategy: Writer AI shifted adoption from technical teams to non-technical business units using natural-language “Playbooks” and reusable “Skills” builders.

• The Lesson: Change happens faster when employees aren’t forced to learn complex code, but can instead dictate workflows using natural language.

  ---

Reimagining the Framework: What AI Change Management + AI Design Thinking Looks Like

Distyl AI’s “Distillery” platform concept, which decomposes written Standard Operating Procedures into auditable AI workflows called “Routines”, is the perfect example of AI Design Thinking in action in my opinion.

• Empathize & Discover (The SOP Audit): Instead of standard change-readiness surveys, map existing human friction points. Have teams upload their written SOPs.

• Define (Decomposing the Workflow): Use AI to automatically break down those manual SOPs into discrete tasks (e.g., validating an order, checking policy). Let the human define the guardrails, not the code.

• Ideate & Prototype (The Visual Sandbox): Replace long training manuals with iterative visual builders. Subject-matter experts (non-technical staff) review, test, and adjust the prompts in real time. This turns them from passive recipients of change into active builders of the system.

• Continuous Change Fitness (The Human-in-the-Loop): Build a continuous learning loop. When a complex exception occurs, the system triggers a human-in-the-loop review. The human’s feedback automatically optimizes the AI model going forward.

• The 80/20 Rule & Bandwidth Creation: When AI agents handle 80% of routine knowledge work, the change manager’s new job is to help teams redefine the human’s role as the critical 20% strategic thinker and exception handler. The narrative shifts from cost-cutting to capability.

Look out for a part two on change management, as this post only scratched the surface on existing models, then and now. There is more to be said on trends and deployment.

💡 See my summer series of six posts to an AI-Native and an AI-Integrated approach. I have used Design Thinking tools & updated existed frameworks to solve a real-world bottlenecks for a life insurance client, mirroring the rapid delivery models used by Distyl and Writer. In matter of days, alongside actuaries we produced vibe coded solutions as business cases ready for deployment.

---

In Closing

Change Management models must intentionally include AI adaption for organizations to perform in an AI-native AI first market. Change is always the hard part, making it exponentially critical to integrate in your growth strategies.

If change managers don’t develop at least a 30% AI mindset, they will not just fail their projects, they will find themselves automated out of the transformation discussion entirely by forward-deployed engineering teams and agentic systems. This is already happening at AI-native startups.

I am curious - where do you sit in this AI transformation?

Did your company invest in training, or are you self taught?

How many tools do you use daily/weekly?

Does your leadership understand and use AI, or do they leave it to teams?

Finally, I want to know

Is your organization still trying to fit AI into a legacy Prosci framework or an endless SaaS POC cycle? Or are you shifting toward co-creating workflows with your teams? Let’s discuss below.

👉 For paid subscribers 👉 Change Management frameworks and cheatsheets with an AI update are included. Design Thinking also plays a role, see the earlier post.

Read more

Finale of ‘From Stuck to Launched: A Humanistic AI Framework for Getting Started.’

Every business book puts market research first.

I put it last, and I want to explain why, because when I first mapped this series, I had it in position two. Right after idea validation. That is where it belongs if you think of research as the thing that tells you whether to proceed. I was writing it, as I teach it, in that same order.

Then I looked at what I actually do with clients. Nobody has ever hired me to tell them whether a market exists. They hire me when they have already started, are already spending, and cannot tell whether any of it is pointed at something real. By the time market research is useful, there is usually a product, a price, and a bill.

So I moved it to the end.

Research before contact with reality is procrastination with a bibliography. You can spend three months producing a document that tells you what a market is worth and still not know whether one person will pay you. I have watched founders commission decks instead of making decisions, and seen leaders presenting beautiful market analyses and no position.

Research after five stages of real customers, real prices, and real math is a different instrument entirely. It verifies. It arms you for the conversation with the board or investors. It tells you what to build next.

This is the last framework in the series. It is also the one that tells you whether the other five were pointed at anything.

A Note on AI

Secondary research became free this year. Twenty-page market analyses in minutes. Competitive landscapes rendered while you make coffee.

Which means everyone’s research now looks the same.

That is the part people have not caught up to. The synthesized report is confident, complete-looking, and close to identical to the one your competitor generated an hour earlier from the same public sources. I wrote about this in my essay on AI slop. Market research is where it costs the most, because a report that reads like everyone else’s produces a strategy that looks like everyone else’s, and nobody notices until the positioning meeting goes flat.

So the edge moved. It is now in the conversations AI cannot have. The observation it cannot do. The flinch across the table when you say a number out loud, which no transcript captures.

AI made market research abundant. Abundant things are not edges.

The Framework: Verify, Do Not Validate-Shop

Research exists to change a decision.

That is the whole test, and most research fails it. If no answer you could possibly find would change what you do next, you are not researching. You are collecting permission.

We will use Jobs-to-Be-Done and Rob Fitzpatrick’s The Mom Test. Sizing is bottom-up throughout, and I will explain why top-down sizing is worse than useless.

Before we start

Four questions this post will make you answer:

• Can you name ten real members of the market you claim to serve?

• What do your prospects do today instead of buying anything at all?

• What did your last ten conversations tell you that you did not already believe?

• What is the date you will look again?

TL;DR: Size bottom-up. Compete against the status quo. Do the primary work. Read behavior rather than reports. End every study in a dated decision.

Stage 1. Size It Honestly

One percent of a fifty billion dollar market.

I have sat through that slide more times than I can count, and I have never once seen anyone in the room ask where the fifty billion came from. Not because they are not smart. Because asking makes you the difficult person, and the number is not really there to be interrogated. It is there to establish that the opportunity is big enough to be worth the meeting.

Top-down sizing starts with a figure nobody can verify and applies a percentage nobody can defend.

Bottom-up is harder. How many buyers exist in the beachhead you chose in Post 3 (see Go-To-Market below). What each one pays at the price you set in Post 4 (linked below). How many of them can one channel actually reach? Multiply those three and you get your serviceable obtainable market (SOM), which is the only one of the three sizing numbers that has anything to do with your next twelve months.

A market you cannot name ten members of is not a market. It is a hope. Or worse delusion.

The AI Move: Build the bottom-up model with AI, then turn it around and ask which five of your assumptions are most likely wrong. Never use it to fetch a topline market size. That figure came from the same public sources your competitor is quoting.

Ask yourself:

• Can I name ten real members of my serviceable obtainable market?

• Is my sizing built from buyers or from percentages?

• Which single assumption, if wrong, collapses the number?

Stage 2. Map the Competition You Actually Have

Your biggest competitor is almost never a company.

It is the spreadsheet someone built four years ago and everyone still uses. The intern who does it manually. The workaround people complain about constantly and never replace. The sentence “we just live with it,” which I have heard in board rooms and distribution centres and on kitchen tables, and which is the most expensive sentence in business.

Founders lose to nothing far more often than they lose to a named rival. And then they build competitive analyses that do not mention it.

The comparison that matters is not features against price. It is jobs. Who else is currently hired for this job, and what would it cost the buyer to fire them? Sometimes the answer is money. More often it is disruption, retraining, and the risk of looking foolish in front of a boss. That last one is priced into more decisions than anyone admits.

The AI Move: Point AI outward at scale. Competitor pricing pages, changelogs, job postings, review-site complaints. This is the language mining from Post 1 (linked below), aimed at the market instead of at your own customers. AI reads at volume. You interpret.

Ask yourself:

• What do my prospects do today instead of buying anything?

• What would firing that actually cost them, in money and in disruption?

• Have I ever written a battle card for doing nothing?

Stage 3. Do the Primary Work

This is the stage people skip, and I understand why. It is slow, it is uncomfortable, and there is now a machine that will produce something that looks like the output without any of the discomfort.

Fitzpatrick’s discipline compresses to three rules. Ask about their life, not your idea. Ask about past behavior, not future intention. Ask for specifics, not opinions.

“Would you use this?” is a lie-generating machine. Not because people are dishonest. Because they are kind, and because nobody can accurately predict their own behavior about a thing that does not exist yet. I have been on both sides of that question, and I have given the polite answer.

You have done this before in this series. Post 1 was interviews about a problem. Post 4 was conversations about price. Same muscle, wider aperture. You are no longer researching one buyer. You are researching how a market behaves.

The AI Move: AI drafts the non-leading interview guide. Afterward it clusters the transcripts and surfaces the phrases that repeat, as in Post 2. The middle part, the human across the table, has no AI move. That is not a gap in the tooling.

Ask yourself:

• Am I asking about past behavior or hypothetical interest?

• What did this person actually do, pay for, or abandon in the last ninety days?

• What surprised me? If nothing surprised me, I was leading.

Stage 4. Read the Signals, Not the Reports

Markets announce themselves in behavior long before they show up in reports.

Job postings tell you who is hiring against the problem. Pricing page changes tell you what incumbents are defending. Community complaint volume tells you where pain is concentrating. Churn direction tells you which way the water is running.

None of that is in a report, because by the time it reaches a report it is a description of the past.

Here is the part worth sitting with, and it is uncomfortable if you have worked anywhere large. Big organizations do not miss shifts because they lack research. They miss them because they have the most research, produced by the most credible people, describing a market that has already moved. The report is not wrong. It is late. And the more rigorous the process that produced it, the later it tends to be.

The AI Move: Run a monthly signals scan. Point AI at the watering holes you identified in Post 3, have it summarize what changed since last month, and read the summary asking one question. What moved?

Ask yourself:

• What behavior, not opinion, shifted in my market this quarter?

• Which incumbent is quietly changing prices?

• Am I reading the market, or reading about the market?

Stage 5. Decide and Date It

Every piece of research ends in a one-page memo with four lines. What we now believe. What would change that belief. What we are doing about it. When we look again.

The last one is the one people skip, and it is the one that does the work. A belief without a review date quietly becomes a fact. Nobody decides to stop questioning it. It just ages into the furniture.

I have watched teams operate for years on a market assumption that was true when someone made it and had not been true for some time. Nobody was negligent. There was simply no date on it, so there was never a moment when checking was anyone’s job.

Put the expiry on the decision. Same discipline as the pricing calendar in Post 4 and the re-run date in Post 5.

The failure mode this kills is perpetual research. One more study is fear with a methodology.

The AI Move: AI drafts the memo from your artifacts. You sign the decision. AI cannot make the decision, and that limitation is the point of the entire series.

Ask yourself:

• What decision is this research serving?

• What answer would change my mind?

• What is the re-check date? A date, not an intention.

Market Research for Writers and Creators

For the writers and side hustlers who have stayed through all six of these.

Size your niche bottom-up. Not how many people are interested in productivity content. Name twenty-five real people who have the problem you write about. If you cannot, the niche is a hope. Same lesson as Post 1: the best niches sound too narrow at first, right up until narrow is the reason anyone subscribes.

Map the competition honestly. Find the five to ten publications your exact reader already reads. Not to imitate. To find the unserved job. Read their comment sections for the question that keeps getting asked and never answered. That question is your next six months.

You are already sitting on primary research and probably ignoring it. Replies. Which posts got forwarded. Which ones brought strangers rather than people who already know you. Ask every new subscriber how they found you, in the welcome email, where it costs you nothing. Your inbox is a dataset nobody else can see.

📌 Use Case: This series was itself a piece of market research, that I ran in public. One caveat I must share: Weeks before the series launch, I moved the publication section from Travel (where I started) to Business on Substack. My views dropped to less than 50%. The algorithm had to be trained, and while I am (finally) back to steadily gaining a few subscribers a week, my experiment was impacted. I decided not to stop, to run the summer series anyways. Have my numbers drastically changed? Honest answer is No.

Traps:

• Researching instead of publishing. The publishing is the research.

• Imitating the leader in your niche rather than finding the gap they left.

• Treating follower counts as market size. That is reach, not demand.

The Founder’s AI Discipline

• Use AI to read the market at scale, not to replace meeting it.

• Use AI to attack your sizing, not to inflate it.

• Use AI to draft the interview guide, not to imagine the answers.

• Use AI to write the memo, never to make the decision.

Same pattern in all six posts. AI compresses what was slow. It cannot do what was hard.

Closing the Series

Six frameworks, in the order reality imposes them. Validate the idea. Prove the fit. Pick the channel. Set the price. Close the math. Verify the market.

One discipline underneath all of it: contact with reality before commitment of resources. Every framework in this series is a different way of finding out sooner.

I should say what I do not know yet.

I do not know whether this series worked. I know what I published, and I know what the numbers say this week, but the thing I was testing was whether frameworks explained plainly, with my own receipts attached, would find the people I wrote them for. That takes longer than six weeks to answer.

Where the scoreboard stands. My numbers are smaller than I hoped. The whole argument for running this in public was to share the real numbers.

The founding window. When this list crosses 300, thirteen founding spots open at a special price, capped, never repeated. The founding members get 1:1 coaching, walk through the six stages on your problem and founder cohort.

👉 Paid subscribers get the full tools library, including the Bottom-Up Sizing worksheet and the Decision Memo, plus everything from the previous five posts.

The Series Map is free for everyone, whether you have read a word of this or not.

The Series Map

What happens next. I will focus on pertinent topics such as change management for AI adoption.

You do not need permission to build. You need evidence, and you now know how to collect all six kinds.

Thank you for joining me in this special six-part summer series: From Stuck to Launched: A Humanistic AI Framework for Getting Started.

Read more

A refresher for the rusty and an easy-to-follow starting point for anyone with revenue in sight — especially now, in the AI era, when software has marginal costs again, and most founders haven’t noticed yet.

Most founders can tell you their revenue. Very few can tell you what they earn or lose on one customer.

I have witnessed founders celebrating growth that was quietly multiplying a loss. I have consulted with teams whose dashboards glowed green while every new customer made the hole deeper. I have watched creators sprint toward audience milestones without once dividing their revenue by their hours.

The pattern is always the same. Not a lack of ambition. Not a bad product. Just no framework for the only question that decides whether growth is a strategy or an accelerant: do you make money on one?

You validated the idea (Post 1). You found fit (Post 2). You picked your channel (Post 3). You set your price (Post 4). Now the math has to close. This is my framework. Explained like a human, not a finance class.

A Note on AI

AI has changed three things about unit economics:

1. Marginal cost is back. For twenty years, software’s dirty secret was that serving one more customer cost almost nothing. AI products pay per token, per inference, per call. Margin is a design decision again, and many founders are designing it by accident.

2. The math got easier to do. AI can build your cohort models, cost breakdowns, and scenario analysis in minutes. There is no longer an excuse for not knowing your numbers.

3. ‘Cheap to build’ now hides ‘expensive to serve’. You can ship something that costs you real money every single day a user loves it in a weekend. The better your retention, the faster you bleed. That inversion is new, and it is catching people. Cheap to build now hides expensive to serve. A user who opens your product five times a day is either your best case study or your biggest liability, and the only way to know which is to have done the math.

I will weave AI moves throughout. Use them to see your costs clearly, not to make the spreadsheet say what you were hoping.

Today we cover Unit Economics.

We anchor in the two ratios operators have used for decades: lifetime value (LTV) against acquisition cost, and the payback period, popularized for startups by David Skok’s SaaS metrics, and we integrate the discipline you already know from this series: real numbers over good feelings.

The Napkin Math worksheet and the Levers one-pager are included for paid subscribers at the end of this post.

The Framework: Make Money on One, Then Multiply

Growth does not fix unit economics. It multiplies them. Also including the bad ones.

The whole discipline reduces to five honest numbers about a single unit:

1. What is one unit? A customer, a seat, an order, a subscriber. Pick the one most relevant to your business.

2. What does one unit pay you? Revenue per unit, over its lifetime.

3. What does one unit cost to serve? Everything it takes to deliver, including the costs you have been calling temporary.

4. What does one unit cost to acquire? Everything it takes to win, including your own hours.

5. Do the ratios close? Value against acquisition, and how fast the cash comes back.

An old operator’s saying covers the stakes: revenue is vanity, margin is sanity, cash is reality.

At the end of this post, you will have answered these key questions

• What do you earn or lose on one customer? It’s a number, not a feeling or guesswork.

• What does one customer actually cost to serve, once your time and your AI bill are in the math?

• How many months until a new customer pays back what they cost to win?

• If you doubled tomorrow, would your economics get better, or would your losses?

TL;DR: Unit economics is the question ‘do I make money on one?’ asked without flinching. Get one unit profitable, or know exactly why it isn’t yet and when it will be, before you invest further. In the AI era, growth is easier than ever to buy, and margin is easier than ever to lose.

↳ Next Step: Tonight, on one page: what one customer pays you, what one costs to serve, what one costs to acquire. Three numbers. If you cannot produce them, that is your finding.

Stage 1. Name Your Unit

Before you can do the math, decide what you are counting.

The unit is the metric of your business: the one thing, when repeated, becomes the whole company. For a SaaS product, it is usually a customer or a seat. For a marketplace, a transaction. For a service business, an engagement. For a newsletter, a subscriber. If this is unclear, stop. Go back to Idea Validation (post 1) and Product Market Fit (PMF) (post 2) and have the business model conversation before you go further.

Red Flag: Founders get this wrong in a revealing way: they pick the unit that flatters them. Counting signups when the business runs on paying seats. Counting downloads when the business runs on subscriptions. The metric or unit you choose determines every number downstream. Therefore, choose the one that maps to money, not the one that maps to applause and vanity metrics.

📌 Use Case: At AARP, the problem was never a shortage of measurement. Leadership had numbers. What they did not have was agreement on which number meant anything.

Reporting showed training sessions delivered, program registrations, and media impressions. Impressions were the worst of the three, because they counted reach rather than impact. All three went up every quarter, and every quarter they made the case for changing nothing.

The members told a different story. Bilingual seniors walked into free member workshops nervous about the device in their pocket, the one thing standing between them and decisions about their family and their own retirement. They walked out able to check a bank balance, pay a bill, and find their own retirement information on the AARP site without asking a grandchild or a neighbor for help, in Vietnamese or Chinese or Spanish.

That was the unit. Not a registration. A member who left able to do it alone.

Then we counted differently. Registrations counted everyone who walked in. The honest unit counted everyone who walked out capable. Same room, same cost, same quarter, two very different numbers. One said the program was working. The other said which parts of it were.

That distinction is what made a serious conversation about where the money went possible.

Teams do not usually measure badly. They measure what is easiest to pull from the system they already have, and then defend it because it flatters them.

The AI Move: Describe your business model to AI in five sentences and ask: what is the unit of this business, and what is the vanity unit I am most likely tracking instead? Then check which one your dashboard actually shows.

Ask yourself:

• What is the smallest repeatable thing that brings in money?

• Does my main dashboard count the unit, or something that merely correlates with it?

• If I had exactly one unit, could I compute my whole P&L from it?

• Am I counting what pays me, or what flatters me?

  ---

Stage 2. Know Your True Cost to Serve

Gross margin is where honest businesses are separated from optimistic ones.

Cost to serve is everything it takes to deliver one unit its value: hosting, tooling, support, payment fees — and in the AI era, inference.

Tokens are the new COGS. A traditional software product could wave at 80%+ gross margins. An AI-heavy product can quietly sit far below that, and the founder often finds out only when usage grows.

Two costs founders systematically exclude, and must not:

• Your own time. The manual onboarding, the concierge support, the I will just handle this one myself. At validation stage, doing things that don’t scale is the right move. We covered this in Post 2. But it is a loan your unit economics are taking out, and Stage 2 is where you write down the balance.

• The AI bill at scale. Price the tokens at your best customer’s usage, not your average user’s. Your most engaged users are your most expensive ones. Design for the margin you want at their consumption. Best customer = ICP super user.

📌 Use Case: I spent a stretch of my SAP years working on a global integrated supply chain for a consumer healthcare retailer. Enormous company. Sophisticated operators. We organized the work around four themes, and the one that kept swallowing the others was visibility. Not analytics. Not prediction. Visibility. Where is my stock right now? What is actually happening inside the fulfillment network? The honest answer, from people who ran that network for a living, was that they could not see it.

Every cost decision in that business was being made against data that did not exist yet. Not wrong data. Absent data. I think about that whenever a founder tells me their margins are fine. The question is never whether the number is good. The question is whether the number is real, or whether it is a shape you have filled in because the alternative is admitting you cannot see.

Scale does not fix this. That company had every resource available and still could not answer where is my stock? You will not outgrow the problem. You will only make it more expensive to discover.

The AI Move: Here is my AI build cost-to-serve model. I spend monthly [Subscription fees and costs: $100/month, 15-20 hours at $150/hr (my best honest rate is $3,100). Here is my unit. Allocate the costs and tell me my real gross margin — then tell me which single cost line grows fastest if I 10x.” The last question is the one that matters.

My cost checklist (based on a coaching model that I have not launched yet):

• What is my gross margin per unit, with my time priced in at what I would pay someone else?

• What does my heaviest user cost me per month, and is my price above it?

• Which cost line scales with usage, and which is fixed?

• If usage doubled tomorrow, does margin improve or evaporate?

  ---

Stage 3. Know Your True Cost to Acquire

Customer Acquisition Cost (CAC) is your total cost of all sales and marketing efforts needed to win a new customer.

CAC is not your ad spend divided by your customers. It is everything it took, divided by your customers.

Fully loaded CAC includes the ads, the tools, the content that took twenty hours to write, the conference ticket, the founder’s selling time at an honest hourly rate. Founders who exclude their own labor from CAC are borrowing from the same place they borrowed in Stage 2 (themselves), and the loan comes due the day they try to hire someone to do what they did for free.

Two disciplines from Post 3 carry straight into the math:

• Compute CAC per channel, not blended. A blended CAC averages your best channel with your worst and tells you nothing actionable. Channel-level CAC tells you where to put the next dollar, and it is the number your 90-day playbook was built to produce.

• Segment it by cohort. Your CAC this quarter, not since founding. Channels decay; early customers were friends-of-friends. Fresh cohorts tell the truth.

Not a Use Case: I get this question constantly: how much should a founder pay themselves? The honest answer is that most founders take home less than their early employees, and they know it, and they have made peace with it.

What they have not made peace with is that the same hours show up somewhere else, unpaid and uncounted. If you spend twenty hours a month selling, that time belongs in your customer acquisition cost. It is not free because no invoice was raised for it. It is deferred, and the bill arrives as the thing you did not build, the customer you did not call, or the year you cannot get back.

The founders who skip this line do not get a wrong CAC. They get a flattering one. And a flattering CAC makes a channel look cheap right up to the moment you try to hire someone to run it, and discover that what you were actually doing was subsidizing it with your own life.

The AI Move: Run your acquisition ledger through AI monthly: Here is everything I spent: money and hours, on acquisition, by channel. Here are the customers each channel produced. Compute channel CAC, and flag anything I appear to be leaving out. The flag is usually your own time. Let it call you on it.

Your acquisition checklist:

• What is my fully loaded CAC, per channel, this quarter?

• What is my time worth in that math, and did I include it?

• Which channel’s CAC is trending up, and do I know why?

• If I removed myself from acquisition tomorrow, what would replacing me cost per customer?

  ---

Stage 4. Run the Two Ratios

Two pieces of arithmetic decide whether your machine deserves fuel.

Ratio 1 — Lifetime value against acquisition cost. Lifetime value is what one unit pays you over its life, at gross margin, not revenue. LTV computed on revenue is the single most common flattering error in founder decks. The operator’s heuristic: you want lifetime value at roughly 3x your acquisition cost or better. Below that, you are paying too much for customers, keeping them too briefly, or earning too little margin while they stay. At 1:1, you are running a very tiring charity. Eg: If your CAC is $1,000, LTV at 3x is $3,000+.

Ratio 2 — Payback period. How many months of gross margin does it take to earn back what the customer cost to acquire? The common bar: under 12 months. This ratio is about cash, and cash is the one you can die of. A business can have beautiful lifetime math and still starve waiting for it.

Treat both numbers as heuristics, not laws. They were built on venture-backed SaaS, and your business may differ. But if you cannot state your two ratios, you are not exempt from them. Perhaps you are uninformed about them.

📌 Use Case: See stage 1 use cases.

The AI Move: Give AI your five numbers from this post and ask for the two ratios plus a sensitivity table: Show me what happens to payback if churn improves 20%, if price rises 15%, if CAC doubles. Ten minutes, and you will know which variable owns your future.

Your ratio checklist:

• What is my LTV: computed on gross margin, not revenue?

• What is my LTV: CAC, per channel?

• How many months to payback, and can my cash survive that many months at my growth rate?

• Which ratio is my binding constraint right now?

Stage 5. Fix the Right Lever, in the Right Order

Bad unit economics have four levers.

Founders habitually reach for the wrong one first.

The levers, in the order that usually pays:

1. Churn. Keeping a customer you already paid to win improves LTV, payback, and CAC efficiency simultaneously. It is the only lever that pulls the other three. (You built the retention instruments in Post 2. This is why.)

2. Price. The fastest margin lever there is, and the one fear postpones longest. You did this work in Post 4. If your ratios fail and nobody has flinched at your price, you know your first move.

3. Cost to serve. Automate the manual, renegotiate the stack, design the AI usage instead of absorbing it. Real gains here, but slower than a price change.

4. CAC. The lever founders reach for first. More content, more ads, more hustle, and usually the weakest, because acquisition improvements decay while the other three compound.

Notice what is not on the list: growth. Growth is not a lever on unit economics. It is a multiplier of whatever your unit economics already are.

📌 Use Case: A first-time founder came to me convinced the problem was acquisition. Cost per customer was too high, so the answer was to get it down. That is where almost everyone starts, because acquisition is the number you see every day.

The leak was churn. Customers were arriving fine and leaving quietly, and every dollar spent on acquiring more of them was pouring into a bucket with a hole in it. Fix the hole first, and the same spend suddenly works. Fix the spend first, and you have made a leaking bucket marginally cheaper to fill.

This is worth sitting with, because it looks like it contradicts the levers above. It does not. Price is the fastest lever when your ratio is thin, and you do not yet know why. Churn is the first lever when you do know why, and the answer is that people are leaving. Diagnosis comes before sequence. The framework is not a running order you follow blindly. It is a way of making sure you have asked which thing is actually broken before you spend three months fixing something else.

After she redirected, retention climbed. Then something she had not been aiming at happened: the customers who stayed started recommending her. The acquisition cost she had spent months trying to attack directly came down on its own, because the people who stick around long enough to get value are the ones who tell someone else. That is organic pull, the fifth PMF signal from Post 2, and the one you cannot manufacture. It arrives as a consequence of something else being right.

She never did fix CAC. She fixed the reason it was high.

The AI Move: Ask AI to referee your instinct: Here are my four levers and my current numbers. I want to work on [name the one you are drawn to]. Argue against me: which lever does the math say to pull first? You are allowed to overrule the math. You are not allowed to not know it. Make informed decisions.

Your lever checklist:

• Which lever am I emotionally drawn to, and which does the math point at?

• What is my churn doing to my LTV, in dollars?

• What would a 15% price increase do to payback? (You have the Post 4 worksheet. Use it.)

• Am I about to scale a machine that loses money per unit?

  ---

Unit Economics for Writers and Creators, a Substack Edition

Your newsletter has unit economics. Ignoring them does not make you an artist; it makes you unpaid.

Your unit is a paid subscriber. Your COGS is your time. The platform takes its percentage, the payment processor takes its cut, but the real cost of goods is the hours each post takes you. That number, which almost no writer computes, is the difference between a publication and an expensive habit.

The napkin math, creator version:

• Revenue per unit: your monthly price × the platform’s share left to you.

• Cost to serve: (hours per post × posts per month × an honest hourly rate) ÷ paid subscribers. Notice what this means: your cost to serve falls as paid subscribers grow — writing is one of the few businesses where the same COGS serves the tenth customer and the ten-thousandth. That is the entire economic argument for being a writer. It only kicks in past the point where most writers quit.

• CAC: mostly your time on distribution — the Notes, the LinkedIn posts, the outreach from Post 3. Price those hours too.

• LTV: monthly margin × average months a paid subscriber stays. Track your paid churn from day one; it is the number that decides whether growth compounds or leaks.

The three honest verdicts the math can return: a business (margin per hour you would accept as a salary), an investment (negative now, with subscriber growth that credibly flips it — most good newsletters live here for a year or more), or a hobby (negative with no trajectory — which is a legitimate thing to choose, as long as you chose it).

📌 Use Case: Mine, since I owe you the same honesty I asked of you. Ninety-one subscribers. Four paid. You have now met my four friends.

My revenue per post is $5.34. My hours per post are ten to twelve, across research, drafting, editing, and the visuals. That works out to roughly fifty cents an hour, which is best described as philanthropic.

Two things I should say plainly rather than let you find them. I raised my prices three weeks ago, from five dollars a month to fifteen. I have no idea yet whether that worked. Four subscribers is not a signal, it is a rounding error with names attached, and I am publishing a post about unit economics while my own key input is three weeks old and untested.

Worth noting that all four subscriptions are annual, which means I collected a year of lifetime value upfront. The list is small. The structure is right. Those are different problems, and only one of them is fixed by writing more.

And my monthly price sits above the eight-to twelve-dollar range I recommended to you in Post 4. That is deliberate. The annual works out to ten a month and founding to twelve and a half, both inside the range. I priced the monthly high because I would rather you took the annual. But I am not going to quietly hope nobody noticed the gap between my advice and my checkout page.

By the framework above, The Fluid Identity is not a business today. It is an investment. The whole summer strategy you have been watching is the argument that the trajectory flips it, and the napkin math gets run again, in public, at three hundred subscribers.

The AI Move: Paste three months of your writing and distribution hours plus your subscriber numbers into AI and ask for your creator P&L: margin per post, effective hourly rate, and the paid-subscriber count at which your rate crosses minimum wage, then your market rate. Knowing those two thresholds changes how the hard weeks feel.

Common creator traps:

• Counting free subscribers as the unit. Free subscribers are pipeline. Paid subscribers are the business.

• Excluding your hours because you enjoy the work. Enjoyment is not a margin.

• Chasing CAC (more platforms, more promotion) when churn is the leak. A canceled paid subscriber costs more than a new free one is worth.

• Never deciding whether this is a business, an investment, or a hobby. All three are honorable. Only one thing is fatal: not knowing which you are running.

  ---

The Founder’s AI Discipline (for Unit Economics)

AI can make your math effortless. It can also make your denial more sophisticated.

• Use AI to compute your real margins, not to build the model that hides them.

• Use AI to price your own time into the math, not to keep donating it invisibly.

• Use AI to stress-test the ratios, not to decorate the deck.

• Use AI to watch your inference costs, not to discover them in an invoice.

The founders who win on unit economics in the AI era are not the ones with the most elegant models. They are the ones who can answer ‘what do you make on one?’ without opening a spreadsheet — because they already know.

Before You Move to the Next Stage

Ask yourself one question: If I doubled tomorrow, would I be twice as strong — or twice as broke? {ouch!}.

Growth feels like the answer to everything. It is the answer to nothing that is broken at the unit level.

The founders I have sat with in their messiest moments funded growth that multiplied a per-unit loss, retention that masked a margin collapse, and a successful launch that cost more per customer than it earned — did not need more customers. They needed five honest numbers and the nerve to act on what the ratios said.

That is what these frameworks are for.

Next week, the series finale: Market Research — how to see the market clearly before it corrects your assumptions.

Tell me which number you want to go deeper on. I will expand it into a full post, worksheet, or coaching session.

Read more

A refresher for the rusty and an easy-to-follow starting point for anyone with paying customers in sight — especially now, in the AI era, when building has never been cheaper, and underpricing has never been easier to justify.

Since this post is a few days late, the Summer Series · Post 5 of 6 will drop soon.

📌 Bonus use case added after publication in Stage 3, below. I wrote this post without a case from my own consulting years, which was an omission. This one had been sitting in my portfolio the whole time, and it is the clearest example I have of the shift this stage is about.

Most founders treat pricing as the last decision. A number to fill in after the product is built, chosen by glancing at competitors and subtracting a little. It is the most expensive glance they will ever take.

I have coached founders who agonized for months over features and set their price in an afternoon. I have consulted with teams whose product was genuinely better than the incumbent’s — priced 40% below it, teaching every prospect that it must not be. I have watched creators give away their best work for years because charging felt presumptuous.

The pattern is always the same. Not a lack of value. Not a bad product. Just no framework for having the one conversation that decides whether the business survives: what is this worth, and to whom?

You validated the idea (Post 1). You found fit (Post 2). You picked your channel (Post 3). Now: what do you charge? This is my framework. Explained like a human, not a spreadsheet.

A Note on AI

AI has changed three things about pricing:

1. Feature moats collapse faster. What took a competitor a year to copy now takes weeks. Pricing anchored to features is pricing anchored to sand.

2. Willingness-to-pay research is faster. AI can draft surveys, role-play objections, and cluster a hundred customer conversations into patterns in an hour.

3. The race to the bottom got seductive. When everyone can build cheaply, underpricing feels safe. It is the opposite: it makes you indistinguishable from the flood.

I will weave AI moves throughout. Use them to hear your customers, never to justify the number you already wanted.

Today we cover Pricing.

We anchor in value-based pricing — the discipline Madhavan Ramanujam codified in Monetizing Innovation: have the price conversation before you finish building, not after — and the Van Westendorp Price Sensitivity Meter, four questions that have quietly guided pricing research for decades. Both integrate with what you already know from this series: talk to real customers before you commit.

Summary visuals and the Pricing Conversation worksheet are included for paid subscribers at the end of this post.

The Framework: Price the Problem, Not the Product

Most founders price based on cost or competitors. Both are ways of avoiding the customer.

Three pricing postures, in ascending order of courage:

1. Cost-plus → arithmetic. What it cost me, plus margin. Ignores the customer entirely.

2. Competitor-based → imitation. What they charge, minus fear. Outsource your strategy to someone else’s.

3. Value-based → strategy. What the solved problem is worth, shared between you and the customer.

Cost-plus is arithmetic. Value-based is a strategy.

At the end of this post, you will have answered these key questions

• What does this problem cost your customer per month — in money, time, or risk?

• Have you asked a single real customer what they would pay? Not guessed. Asked.

• Does your pricing metric grow when your customer’s value grows?

• When did you last raise prices — and if never, what are you afraid of finding out?

TL;DR: Price is a positioning statement with a number. Anchor it to the customer’s problem, test it in real conversations, set it higher than feels comfortable, and revisit it like the living decision it is. In the AI era, building is cheap — underpricing is the fastest way to look like everyone else.

↳ Next Step: Before your next build sprint, run one willingness-to-pay conversation using the four questions in Stage 3. One. This week.

Stage 1. Find the Cost of the Unsolved Problem

Before you name a price, name what the problem costs.

Every strong price rests on one number the founder can say out loud: what it costs the customer to not buy — hours lost per week. Revenue leaking per month. Risk carried per quarter. If you cannot quantify the unsolved problem, you are not ready to price the solution — you are ready for more customer conversations.

Once you have that number, pricing stops being a guess and becomes a fraction: you are asking for a share of the value you create. A tenth of a $50,000 problem is an easy conversation. A $500 price with no problem attached to it is a hard one.

📌 Use Case: When we ran the caregiving experiment, the adjacent insight was on the provider side. Empty appointment slots were cutting physicians' daily productivity by 12.5% to 14%, for a smaller practice, that landed somewhere between $50,000 and $150,000 a year. That number changed the conversation entirely. We were no longer discussing whether a scheduling tool with a senior safe ride option was worth building. We were discussing what share of $100,000 a year it was reasonable to ask for. This led to a conversation with Lyft’s strategic unit.

The AI Move: Use AI to build the cost-of-problem model with you: “My customer is [how]. The problem is [what]. Help me estimate what this costs them monthly in time, money, and risk. List the three assumptions most likely to be wrong.” Then verify those assumptions in real customer conversations. AI drafts the model. The customers correct it.

Ask yourself:

• In one sentence, can I state what the unsolved problem costs my customer, with a number in it?

• Is that number theirs (from conversations) or mine (from a spreadsheet or extrapolation)?

• What share of that value am I asking for (the one used to determine SOM)?

• Would my customer grasp the market-sizing metrics, or scoff at it?

  ---

Stage 2. Choose What You Charge For

The unit you price on matters more than the number on it.

Per seat. Per use. Per outcome. Flat monthly. One-time. The pricing metric is the exchange rate between your customer’s success and your revenue — and the test of a good one is simple: when your customer gets more value, do you earn more? If they can grow 10x on your product while paying you the same, you chose the wrong unit. If they get charged more while getting nothing more, so did you — they will feel it, and they will leave.

📌 Use Case: For Grow with Scaleups' new program launch, we led with market research and competitors’ offering a 6-month free runway and cloud credits. Our internal approvals were more challenging than the market discussions. We discovered that high-growth startups were interested in talking to SAP and accessing its customers and partner base.

The AI Move: Ask AI to generate five different pricing metrics for your product and stress-test each: “For each, describe a customer who wins unfairly and a customer who gets punished unfairly.” The metric with the fewest unfair edges wins.

Your metric checklist:

• Does my revenue grow when my customer’s value grows?

• Can a customer predict their bill without a spreadsheet?

• Does the metric punish the behavior I want to encourage? (Per-seat pricing punishes adoption. Know the tradeoff.)

• Am I charging for what is easy to count, or for what actually tracks value?

  ---

Stage 3. Ask Before You Set

You do not find your price in a spreadsheet. You find it in a conversation you are avoiding.

The Van Westendorp method is four questions, asked of real potential customers:

1. At what price would this be so expensive you would not consider it?

2. At what price would it feel expensive, but worth considering?

3. At what price would it feel like a bargain?

4. At what price would it be so cheap you would question the quality?

Between the answers sits your acceptable range — and question 4 is the one founders forget exists. There is a price below which you are not attractive; you are suspicious.

The harder, better version: at the end of a customer conversation, name a specific price and watch the reaction. Not “would you pay for this?” — everyone says yes to hypotheticals. Ask: “This will be $X a month. What would you expect it to include?” The question assumes the purchase and surfaces the real objections. Silence, flinching, immediate negotiation, or (the signal you want) questions about onboarding: each tells you something a survey never will.

📌 Use Case: You might remember we had created a ‘fake website’ demand test in Post 2. We collected emails in 36 hours and got executive greenlight to build a caregiving services option that supported caregivers. Family caregiving in the U.S. reaches an economic value of $1.01 trillion annually, driven by 59 million caregivers providing 49.5 billion hours of unpaid labor. This is valued at an average of $20.41 per hour, and a jaw-dropping workforce equivalent that matches 23.8 million full-time jobs, or 17% of the U.S. full-time workforce (source: AARP).

📌 BONUS Use Case: During my SAP years, I was brought into a strategic relationship with a supermajor energy client that had gone wrong in a specific way. The account team was seen as a vendor. The client leadership dictated what should be done, and had written the governing document themselves.

That is the purest form of deliverable-based work. The client defines the scope, you execute the list, and the value you can capture is capped at the value of the tasks on it—no amount of doing the list better changes that ceiling.

We spent ninety days on the reset. Most of it was one-to-one interviews with their top leadership, and most of those conversations were uncomfortable, because the useful ones went straight at the thing nobody had named: what kind of relationship this actually was. What came out of it was a different document. Behaviors and outcomes, with responsibility assigned, aligned to their leadership behaviors and their strategy rather than to a task list.

The relationship stopped being about what we would deliver and started being about what would change. That is when the serious conversations became possible.

This is the shift most professional services people are stuck on one rung below. Hourly, then engagement, then outcome. Each step up is a change in what is being bought, not what is being charged, and the price follows the unit rather than the other way around.

The AI Move: Before real conversations, have AI role-play your three toughest customer personas responding to your price. Practice the flinch. Then remember what I told you in Post 1: synthetic customers sharpen your questions; they never replace the answers. AI cannot fake a real person reaching for a real credit card.

Your conversation checklist:

• Have I asked the four questions of at least five real potential customers?

• Have I named a specific number out loud to a real prospect and sat in the silence after?

• What did the “so cheap it’s suspicious” threshold turn out to be?

• Am I pricing from data, or from my own fear dressed as market research?

  ---

Stage 4. Set the Number Higher Than Feels Comfortable

If nobody flinches, you are underpriced.

The uncomfortable truths about the number itself:

• Almost every founder underprices. The fear of hearing “no” prices the product, not the value. A price that produces zero pushback is not good; it is money and positioning left on the table.

• Your price talks. Before a single feature is used, the number tells the customer what category you are in. Discount positioning attracts discount customers — the ones who churn fastest and demand the most.

• Three tiers beat one price. A high anchor tier makes the middle tier feel reasonable; a floor tier catches the cautious. Most buyers take the middle. Design it that way on purpose.

• Raising is easier than you fear. Grandfather existing customers, give honest notice, attach the raise to real added value. The customers you lose at a fair price increase were the ones churn was coming for anyway.

📌 Use Case: A clinical ops startup founder has underpriced the platform that drastically cuts the administrative load and leakage of new and returning patients. Initially, this makes sense for an unknown platform and founder who wants to build a trusted customer base and drive numbers. We discussed raising prices and surviving it. Her fear is real. My advice was to track the number after (the churn that didn’t occur, and the uptick in revenue). This is the stage where a real number changes minds, and the fear of raising prices is universal and only receipts cure it. Another founder move I have observed is to use third-party sellers who build your pipeline.

The AI Move: Model three-tier structures with AI and pressure-test the anchoring: “Here are my three tiers. Which would each of my personas choose, and what would each wish the middle tier included?” Then check the model against the willingness-to-pay conversations from Stage 3. The model proposes. The conversations dispose.

Your number checklist:

• Does my top tier make my middle tier look sane?

• When did a prospect last push back on price? (If never: raise it.)

• Is my price a category statement I am proud of, or an apology?

• What exactly would I add or change to justify 25% more?

  ---

Stage 5. Treat Pricing as a Process, Not an Event

The price you launch with is a hypothesis. Most founders never run the second experiment.

Pricing discipline, on a calendar:

• Quarterly: Review win/loss against price. Where did price kill the deal? Where did nobody blink?

• Twice a year: Re-run the value math from Stage 1. Your product improved. Did the customer’s problem grow? Then the value you deliver did too.

• Annually: Raise, restructure, or deliberately hold — but decide. An unchanged price for two years is not stability; it is drift.

• Always: Resist the discount reflex. Discounts train customers to wait. Add value instead — or shorten the commitment, never the price.

📌 Use Case: I will defer to working in SAP’s Customer Innovation practice. We did not articulate updates in product and business value. This is common in large tech firms where client pricing is calcified for years. Revisiting this as a discipline and running experiments on added value, new capabilities, and opportunity is relevant for both startups and large incumbents.

The AI Move: Set a quarterly ritual: paste your win/loss notes, churn reasons, and support themes into AI and ask, “What is the evidence my price is too low, too high, or measuring the wrong thing?” Fifteen minutes, four times a year — the same cadence discipline as your Post 3 Friday review.

Your process checklist:

• When is my next scheduled pricing review? (A date, not an intention.)

• What evidence would tell me I am underpriced? Overpriced?

• What did my last three lost deals actually say about price?

• Have I ever raised prices? What am I waiting to find out?

  ---

Pricing for Writers and Creators, a Substack Edition

Your newsletter has a price whether you set one or not. Free is a price. So is silence about your paid tier. Here is the framework, translated.

The $5 trap. The most common creator pricing mistake is defaulting to the platform minimum because it feels modest. But conversion at $5 and at $10 is not dramatically different — the decision to pay is about conviction, not spare change — which means the $5 price roughly halves your revenue without meaningfully lowering the barrier. Below a certain price, you are not more accessible. You are less credible. (Stage 3’s fourth question — so cheap you question the quality — applies to newsletters with full force.)

What you are actually pricing. Readers do not pay for volume; they pay for one of three things: tools they will use (worksheets, templates, teardowns), access they cannot get elsewhere (community, office hours, replies), or identity (supporting work they believe in — this is what founding tiers really sell). Know which of the three your paid tier is, and say it plainly on the upgrade page.

The founding tier is a willingness-to-pay test, not a revenue plan. Priced meaningfully above the annual rate, capped, and offered with a deadline, it is Stage 3 in miniature: real people, real money, real signal. Five founding members you do not know personally tells you more than five hundred free signups. (I set this exact bar in Post 1 — and if you read the interlude, you know how my own test went, and what I am doing about it.)

Annual pricing: the convention is roughly two months free. Take it; annual subscribers churn less and fund your patience.

Never discount the price. Adjust the value. A newsletter that goes on sale teaches readers that the price is fiction. If conversion stalls, improve the paid offer or grow the free list — the two honest levers.

📌 Use Case: My own pricing decision is being made in public: founding member spots open when this list crosses 300, and will be capped at 20, and will not be repeated. By the time you read Post 6, you will know whether my own framework survived contact with my own readers.

The AI Move: Paste your paid-tier pitch into AI and ask: Which of the three — tools, access, identity — is this actually selling? Where is it hedging? A pitch that sells all three usually sells none.

Common creator traps:

• Pricing at the minimum out of modesty. Modesty is not a pricing strategy.

• Paywalling the thinking and giving away the tools — backward. Thinking earns trust for free; tools convert it.

• Discounting to convert. You will win subscribers who leave at full price.

• Never stating what paid gets. Vague “support my work” converts only the people who already love you — and you have met my three friends.

  ---

The Founder’s AI Discipline (for Pricing)

AI can sharpen your pricing. It can also launder your fear into a spreadsheet.

• Use AI to model willingness to pay, not to hide behind average competitor prices.

• Use AI to rehearse the price conversation, not to avoid having it.

• Use AI to find the value metric, not to inflate the feature list.

• Use AI to test the raise, not to postpone it.

The founders who price well in the AI era are not the ones with the most sophisticated models. They are the ones willing to say a number out loud to a real customer and stay quiet long enough to hear the answer.

Before You Move to the Next Stage

Ask yourself one question: Is my price a decision I made, or a fear I formatted?

Cost-plus feels safe. Competitor-minus feels reasonable. Both are ways of never asking the customer the only question that matters.

The founders I have sat with in their messiest moments — great products priced like apologies, raises postponed for years, discounts that trained customers to wait — did not need better spreadsheets. They needed the nerve to have one honest conversation about value and a framework for what to do with the answer.

That is what these frameworks are for.

Next week: Unit Economics — where your price meets your costs, and the math tells you whether you have a business or a hobby.

Tell me which stage you want to go deeper on. I will expand it into a full post, worksheet, or coaching session.

Read more

A refresher for the rusty and an easy-to-follow starting point for anyone with a validated product, especially now, in the AI era, when reaching people has never been cheaper — and being ignored has never been easier.

Most founders think Go-To-Market is marketing. It is not. GTM is choosing where to show up, for whom, and saying no to everywhere else.

I have coached founders with validated ideas and working products, stuck at zero customers. I have consulted with teams running five channels at once, none of them well. I have watched creators post everywhere, inconsistently, and wonder why nothing compounds.

The pattern is always the same. Not a lack of effort. Not a bad product. Just scattered effort with no framework for focus.

You validated the idea (Post 1). You found an early fit (Post 2). Now comes the part nobody teaches: distribution. This is my framework. Explained like a human, not a marketing textbook.

A Note on AI

AI has changed three things about GTM.

1. Outreach is cheaper. Personalized cold email at scale used to take a team. Now it takes an afternoon.

2. Content is faster. One post becomes ten assets in an hour.

3. Noise is louder. Everyone has the same tools, which means generic outreach is dead on arrival.

The winners in the AI era are not louder. They are more specific. I will show you where AI compresses the work — and where it quietly destroys your credibility.

Today we cover Go-To-Market.

We anchor in Geoffrey Moore’s beachhead strategy from Crossing the Chasm and in April Dunford’s positioning work, integrating the design thinking principle you already know from this series: observe before you act. You do not need to remember the expert names, follow the methodology as I break it down for you.

Summary visuals and the Channel Decision worksheet are included for paid subscribers at the end of this post.

The GTM Core: Mastering Moore’s Technology Adoption Curve (homage to Geoffrey Moore)

Every revolutionary product faces a critical challenge: crossing the chasm between visionary experimentation and mass-market pragmatism.

According to Geoffrey Moore’s foundational theory, a new technology does not launch into a monolith. Instead, it must navigate the technology adoption curve, segmenting potential customers into five distinct categories based on their risk tolerance and purchase motivation. For GTM marketers, success requires recognizing that each category demands a unique marketing strategy tailored to its specific needs, concerns, and tipping points.

The market splits are approximately:

• Innovators (Tech Enthusiasts 3.5%): Driven by the “new.” They accept bugs and incomplete solutions just to be first. (GTM: Provide access and foster a community).

• Early Adopters (Visionaries 12.5%): Motivated by competitive advantage. They will piece together a solution to achieve strategic gains. (GTM: Shift messaging from features to strategic impact).

• Early Majority (Pragmatists 34%): Risk-averse and practical. They demand a whole product, market leadership, and proof of stability before buying. (GTM: Focus on support, ecosystem, and validation).

• Late Majority (Conservatives 34%): Skeptical of change. They only adopt when the product is standardized and fully commoditized. (GTM: Lower the barrier to entry and offer simplicity).

• Laggards (Skeptics 16%): Actively resist new tech. They often block adoption rather than participate. (GTM: Design solutions that bypass their potential to obstruct).

The ultimate GTM victory is sequentially conquering each segment, knowing when and how to shift your marketing playbook as the adoption curve progresses.

The Framework: One Channel, Fully Committed

Most founders try every channel at once. That is a vanity strategy.

The core principle: do not do all the channels. Do ONE channel really well until it works. Then, and only then, add a second.

The four channels:

1. Direct Sales → B2B, high-ticket products and services ($2K+)

2. Content & SEO → customers who actively search for solutions

3. Communities → B2C and trust-based products

4. Partnerships → when another company already has your customers

One channel at 100% beats five channels at 20%. Every time.

At the end of this post, you will have answered these key questions

• Where do your customers actually spend time — not where you wish they did?

• Which channel matches YOUR skill: selling, writing, community, or partnerships?

• Can you commit to one channel for 90 days without flinching?

• Do you know your customer acquisition cost, or are you guessing?

TL;DR: Pick one channel. Commit for 90 days. Measure. Scale or switch. In the AI era, everyone can do everything. Winners choose.

↳ Next Step: Pick your channel using the decision guide below. Block 90 days on your calendar. Start the clock.

Stage 1. Position Before You Promote

Before you pick a channel, know what you are best at and for whom.

April Dunford’s insight has saved more GTM budgets than any ad platform: most GTM failures are positioning failures in disguise. You cannot choose where to show up until you know what you are showing up as.

Positioning is three sentences:

• For [specific customer]

• Who [specific trigger or pain]

• We are the only [category] that [specific differentiated value]

If you cannot fill in those blanks specifically, no channel will save you. You will simply be ignored in more places at once.

📌 Use Case: Remember the young founder from Post 1, the one targeting matchmakers? Here is the part of his story that belongs in this post. He was not building AI for dating. He was building for professional matchmakers, segmented by category, and he could name their trigger in one sentence. That is why his outreach worked on people like me. Before he narrowed it down, his cold outreach was getting nominal responses. After — when the first line of his email told each matchmaker he understood her category, her bottleneck — his reply rate roughly tripled, and he booked interviews with the top names in the space. His product had not changed. His positioning had. Good positioning does the work of a warm intro.

The AI Move: Draft five versions of your positioning statement using AI, each tailored to a different customer segment. Then test them in real conversations. The one that makes people lean in and say “that is me” wins. AI generates the options. Humans pick the winner.

Ask yourself:

• Can I name my exact customer in one sentence?

• What am I the only one or the best at, for them specifically?

• Would my last three customers describe me the same way?

• Am I positioned for a segment, or for everyone?

  ---

Stage 2. Find Where Your Customers Already Are

Not where you wish they were. Where they actually spend time.

This is design thinking applied to distribution: observe before you act. Founders assume their customers live on LinkedIn because the founders do. Then they discover their buyers make decisions at industry conferences, in private Slack groups, or through peer referrals no algorithm can see.

Geoffrey Moore called this the beachhead: one narrow market segment, completely dominated before expansion. Your beachhead is not just who. It is where.

📌 Use Case: During the caregiving experiment at AARP (the one from Posts 1 and 2) we assumed our users lived where every product team assumes they live: app stores, search, Facebook caregiver groups. Two weeks of observation told a different story. The caregivers actually making decisions found solutions through care coordinators, physician front desks, community, and, for our target audiences specifically, faith organizations — channels no analytics dashboard was tracking. The provider side confirmed it: the practices losing $50,000 to $150,000 a year to empty appointment slots were not searching for software. They were asking peers. When we moved our conversations to those faith organizations, interest jumped. The beachhead was not who we thought. It was where.

The AI Move: Use AI to map your customer’s watering holes. Ask: “Where does a [job title] at a [company type] discuss [problem]? Which communities, newsletters, events, podcasts?” Then verify by showing up and listening for two weeks before you pitch anything. AI drafts the map. Only observation confirms it.

Your observation checklist:

• Where did my last three customers first hear about solutions like mine?

• Which three communities or gatherings hold the most of my exact customers?

• Am I choosing this channel because customers are there, or because I am comfortable there?

• Have I listened to this channel before selling in it?

  ---

Stage 3. Pick One Channel (And Say No to the Rest)

Choosing what NOT to do is your first strategic decision. You heard that in Post 1. Here is where it costs you something real.

The honest tradeoffs:

• Direct Sales: $0–500/month. 3–6 months to results. Requires comfort with rejection — reply rates of 3–5% are normal and fine.

• Content & SEO: $0–300/month. 3–12 months to compound. Slowest to start, most durable once it works.

• Communities: $0–500/month. 2–4 months. Requires genuine helpfulness. Converts best because trust is built in.

• Partnerships: Typically 20–50% commission. 1–3 months. Fastest-to-first revenue with business development skills.

Match the channel to three things: where your customers are (Stage 2), what you are genuinely good at, and how much time and money you actually have.

📌 Use Case: I am running this playbook on this publication right now, with just under 100 subscribers as I write this. The temptation is to be everywhere — Notes, LinkedIn, Instagram, podcasts, swaps, all at once. Instead, I have committed to two moves for the next 90 days: direct, personal outreach to people I have actually worked with, and repurposing every post into LinkedIn content, because that is where the leaders and founders I write for spend their time. Everything else is a no until the clock runs out. You will see the results in this series — including the parts that fail. That is a future post.

The AI Move: AI makes it possible to run all four channels badly. Resist. Use AI to go deeper on one: research every prospect before outreach, repurpose one piece of content five ways, monitor community conversations daily. Depth, not breadth.

Your commitment checklist:

• Which single channel best matches my customer, my skill, and my resources?

• What am I explicitly saying no to for the next 90 days?

• Have I written down my weekly volume target for this channel?

• Who will hold me accountable when I want to switch in week three?

  ---

Stage 4. Run the 90-Day Playbook

Commitment without measurement is just stubbornness. Here is the cadence.

Month 1: Commit and launch.

• Direct sales: send 50 personalized cold emails

• Content: publish 2 in-depth posts (2,000+ words)

• Communities: answer 5 questions daily, sell nothing

• Partnerships: approach 5 potential partners

• Target: 1–3 real customer conversations started

Month 2: Measure and learn.

• Track conversion at every step of your funnel

• Identify the message that gets replies. Kill the ones that do not.

• Close 1–5 paying customers

• Calculate your real CAC — including your time

Month 3: Scale or switch.

• If the channel works: 3–5x your volume

• If it fails, switch completely. Do not tweak endlessly.

• The wrong channel does not mean the wrong product.

• Target: 10–20 customers and a repeatable process

📌 Use case: I do not have a client story for this one yet. I have something better: I am on day 21 of my own 90-day clock with this publication. The numbers, including the ugly ones, are coming in a future post.

The AI Move: Run a weekly GTM review with AI as your ops analyst. Paste in your numbers: what went out, what got replies, what converted, what stalled. Ask: “Where is the biggest drop-off, and what would you test first?” Fifteen minutes, every Friday.

Your playbook checklist:

• Did I hit my volume target this week — yes or no?

• What is my conversion rate at each step in the funnel?

• What did I kill this month?

• Am I at day 90 yet? If not, no switching.

  ---

Stage 5. Earn the Second Channel

Most founders add a second channel too early. The second channel is earned, not chosen.

You have earned it when:

• Channel one produces customers predictably, without heroics

• Your CAC is known and stable

• The playbook is documented, and someone else could run it

• Growth is limited by channel capacity, not by broken conversion

You have not earned it when:

• Channel one works only when you personally grind

• You are bored and want something new

• A competitor launched somewhere, and you feel behind

Boredom is not a strategy. Fear is not a strategy. Repeatability is the only permission slip.

📌 Use Case: The founder from Post 2, the one whose organic signups reached 40% by month three, is the cleanest example I have of an earned second channel. We did not add community because paid acquisition got boring. We added it because the paid channel had a documented playbook, a stable CAC, and data showing that organic users were retained better than paid users. The first channel earned its successor. When we shifted budget, CAC dropped by 60% in a quarter, not because community is magic, but because we expanded from strength with a playbook, rather than from restlessness. The 60% was not the win. The discipline that preceded it was.

The AI Move: Use AI to document your channel playbook as you go: the exact outreach sequences, the posts that converted, the objections and answers. When you finally hire or delegate, the playbook is already written. Founders who skip this rebuild from memory lose months.

Your expansion checklist:

• Can someone else run channel one using only my documentation?

• Is my CAC stable over the last two months?

• Is growth limited by channel capacity or by my conversion?

• Am I adding a channel from strength, or from restlessness?

  ---

GTM for Writers and Creators, a Substack Edition

You are not selling software. But you are running a GTM motion, whether you admit it or not. Your newsletter needs distribution just as much as any startup. The channels are different. The principle is identical: pick one, commit for 90 days, measure, then scale or switch.

The four creator channels:

Channel 1: Substack Notes

• Post short, standalone ideas — not promos for your posts

• Engage genuinely on other writers’ Notes in your niche, daily

• What good looks like: steady single-digit weekly subscriber growth that compounds. Notes rewards consistency over cleverness.

Channel 2: LinkedIn

• Post 2–3x per week. Full excerpts, native text — not bare links. One idea per post.

• Comment meaningfully on 5 posts a day in your niche.

• What good looks like: your professional network converting first, then second-degree strangers. If your readers are professionals, this is likely your beachhead.

Channel 3: Writer Recommendations

• The highest-leverage channel on Substack below 10,000 subscribers

• Find 5 writers with overlapping audiences — peers at your scale, not celebrities

• Engage genuinely for a few weeks, then propose a mutual recommendation

• What good looks like: your steadiest passive source. Results vary widely with the partner’s size and fit — a well-matched swap can outperform a month of Notes.

Channel 4: Podcast Appearances

• Identify 10 podcasts that serve your exact reader

• Pitch a specific angle, not a general bio

• Always name your newsletter as the free resource

• What good looks like: lumpy but high-trust growth. One right-fit appearance can beat everything else that month.

The Creator Roadmap (revised — read this carefully):

• 0–100: Seed. Friends, network, direct personal outreach. One-to-one beats one-to-many at this stage. Nothing scales yet, and it should not.

• 100–500: Traction. One channel, fully committed. This is also where a capped founding-member tier belongs — not as a revenue play, but as your willingness-to-pay test from Post 1. Ten real founding members from strangers tell you more than a thousand free signups.

• 500–1,000: Momentum. Add recommendations and podcasts — the second channel you have now earned.

• 1,000+: The fully paid launch. Announced, priced, with a deadline. A launch, not a toggle.

Notice the distinction.

The founding tier early is an experiment. The paid launch later is a product. Founders who confuse the two either monetize an audience that does not yet exist or wait for permission they have already earned.

The AI Move: Repurpose one newsletter into 10 assets: 5 Notes, 3 LinkedIn posts, 1 thread, and 1 podcast pitch. AI does the reformatting. You do the thinking.

Common creator traps:

• Posting everywhere, inconsistently. Compounding requires repetition in one place.

• Chasing virality over subscribers. A viral post that converts nobody is entertainment, not GTM.

• Skipping recommendations because asking feels awkward. It is the single highest-leverage move on the platform.

  ---

The Founder’s AI Discipline (for GTM)

AI can amplify your distribution. It can also automate your irrelevance.

• Use AI to personalize outreach, not to spam faster.

• Use AI to go deep on one channel, not shallow on five.

• Use AI to analyze your funnel, not to inflate your numbers.

• Use AI to document the playbook, not to skip building one.

The founders winning at GTM in the AI era are not the ones sending the most messages. They are the ones whose messages could not have been written to anyone else.

Before You Move to the Next Stage

Ask yourself one question: Am I doing five channels at 20%, or one channel at 100%?

Scattered effort feels productive. Focused effort converts.

The founders I have sat with in their messiest moments — a great product with zero distribution, a launch that reached everyone and converted no one — did not need more channels. They needed the discipline to choose one and the patience to let it compound.

That is what these frameworks are for.

Next week: Pricing — the decision founders avoid the longest and pay the most for.

Tell me which channel you want to go deeper on. I will expand it into a full post, worksheet, or coaching session.

For paid subscribers 👉 The Channel Decision Scorecard, the 90-Day Playbook worksheet, and both one-page visuals are below the line. These are the working tools I use with founders.

Read more

A refresher for the rusty and an easy-to-follow starting point for anyone with a product, especially now, in the AI era, when the cost of building has collapsed, but the cost of building the wrong thing has skyrocketed.

Most founders think they have Product-Market Fit before they do. They confuse traction with truth. In the AI era, this is more dangerous than ever, because building has never been easier, but building the wrong thing has never been more expensive.

Last week, I broke down Idea Validation. Today, we move to the harder question: once you have built something, how do you know if you have found Product-Market Fit (PMF)?

I have coached founders who thought they had PMF because 50 people signed up on Product Hunt. I have consulted with startups that had 200 users but zero retention. I have seen the same pattern in enterprises: a successful pilot with 5 engaged teams, and 20 more that quietly opted out. This was happening at large enterprises like SAP, when it was late to the cloud revolution.

The pattern is always the same. Not a lack of interest. Not a bad product. Just no clarity on what PMF actually looks like.

And now there is a new pattern: founders shipping fast with AI, watching signup numbers climb, and mistaking momentum for fit.

This is my framework. Explained like a human, not a boring dashboard.

A Note on AI

AI has changed three things about how we find PMF:

1. Prototyping is faster. You can ship 3 versions in the time it used to take to build 1.

2. Measuring is faster. AI can analyze user behavior, surface patterns, and flag churn signals before you see them.

3. Faking PMF is easier. AI-generated buzz, synthetic traffic, and vanity metrics can mask the truth.

I will weave AI moves throughout this framework. Use them to see more clearly, not to see what you want to see.

Today we cover Product-Market Fit.

We will use Sean Ellis’s PMF Survey (the gold standard) and integrate it with Eric Ries’s Build-Measure-Learn loop and reference iterative prototyping from design thinking.

Summary visuals for an easy-to-follow framework approach are included for paid subscribers.

The Framework: 5 Signals of Real PMF

Most founders measure signup numbers. That is a vanity metric.

Real PMF has five signals. Learn them. Track them. Do not confuse them with each other.

1. The Sean Ellis Test → Would 40%+ of users be very disappointed without your product?

2. Day 7 Retention → Do 15%+ come back after a week?

3. NPS → Would 40+ of your users recommend you?

4. Willingness to Pay → Do 40%+ of trial users convert to paid?

5. Organic Growth → Are users bringing other users without you asking?

Hit all five, you have PMF. Hit three, you have a partial fit. Hit one, you have interest, not fit.

Use AI as a thinking partner, not a replacement for customer truth.

At the end of this post, you will have answered these key questions

• Would your customers be very disappointed if you disappeared tomorrow?

• Are your users coming back, or just signing up?

• Would you personally recommend your product to a stranger?

• Is your product creating pull, or are you still pushing?

TL;DR: PMF is not a feeling. It is a set of measurable signals. In the AI era, signup counts are cheap. Retention is the new truth.

↳ Next Step: Pick your product. Run it through the 5 signals. If 3+ are strong, double down. If not, iterate on the weakest signal.

Stage 1. Define What ‘Fit’ Actually Means

Before you measure PMF, define what you are measuring.

Most founders skip this. They assume PMF is when growth feels good. That is not a definition. That is a feeling.

Sean Ellis, who coined the term Growth Hacking, offered the sharpest test: If your product disappeared tomorrow, how many users would be very disappointed?

• If 40%+ say very disappointed → You have PMF.

• If 20-40% → You have a partial fit. Iterate.

• If under 20% → You do not have PMF. Do not scale yet.

📌 Use Case: Working with a leadership team rolling out a new internal platform. They celebrated 800 signups in month one. When we ran a very disappointed survey, only 12% said yes. We paused the enterprise rollout, re-scoped the product for the 12% who cared, and doubled fit within 90 days. Signup counts lied. The survey did not. At another company, a ‘fake’ website offering a solution was rolled out. In 36 hours, we saw traffic with people leaving emails to be notified on launch. The executive team gave the go-ahead on the investment.

The AI Move: Use AI to draft your Sean Ellis survey questions in the language your users actually use, not consultant-speak. Then use AI to cluster open-ended responses. Look for the phrases that repeat. That is the language of your PMF.

Ask yourself:

• What is the one job my product is being hired to do?

• Who are the users I most want to hear are very disappointed?

• Am I measuring interest or dependence?

• Would I bet my next quarter on this product surviving?

Stage 2. Track Retention, Not Signups

Signups tell you your marketing works. Retention tells you your product works.

The single most important signal of PMF is: do users come back without you nudging them?

The metrics that matter:

• Day 1 Retention: 30%+ is good. Below 10% is a product problem.

• Day 7 Retention: 15%+ is a PMF signal. Below 5% means no PMF.

• Day 30 Retention: 10%+ means sticky. Below 3% means churn.

If Day 7 retention is under 15%, no amount of marketing will save you.

📌 Use Case: A founder building a personal health marketplace had almost 200 signups in month one. Day 1 retention was 5%. Instead of buying more traffic, we stripped the product down to one thing: matching clients to trainers. Booking, payment, and messaging came later, all manual. Two weeks after relaunch, Day 1 retention hit 35%. Day 7 hit 18%. That is when we knew.

The AI Move: Use AI to run cohort analysis on your user data. Ask: Which cohorts retained best? What did they do in their first session that non-retained users did not? That behavior is your activation moment. Optimize onboarding to get every user to that moment.

Your retention checklist:

• Do I know my Day 1, Day 7, and Day 30 retention numbers?

• Have I identified the “activation moment” that predicts retention?

• Am I obsessing over signups or over returning users?

• What would double my Day 7 retention next month?

Stage 3. Build in Small Loops, Not Big Launches

The founders with PMF are running Build-Measure-Learn loops every week. The founders without PMF are running quarterly launches.

Eric Ries called this the Lean Startup loop. Design thinking calls it iterative prototyping. Same idea: you learn fastest when you release small, watch, and adjust.

Loop cadence for PMF hunting:

• Weekly: Ship one small change. Measure impact. Learn.

• Monthly: Review retention cohorts. Identify what is working.

• Quarterly: Reset your hypothesis. Are you still chasing the same PMF signal?

📌 Use Case: During a caregiving experiment at AARP, we designed a scheduling tool for elder-care visits. Version 1 had 9 features. Retention was 8%. We killed 6 features and left three: match a caregiver, book a slot, get a reminder. Retention hit 34%. The lesson was not less is more. The lesson was: the loop cadence taught us what mattered in weeks, not quarters.

The AI Move: Use AI to summarize weekly user feedback into themes. What did users most complain about, ask for, or praise? Feed that into your next loop. AI is a fast reader. Use it that way.

Your loop checklist:

• Am I shipping something small every week?

• Am I measuring impact on retention, not just engagement?

• Am I killing features that do not earn their keep?

• What is the one hypothesis I am testing this week?

Stage 4. Listen for the Words Users Repeat

The clearest PMF signal is language.

When you have PMF, users describe your product in the same words to their friends. They repeat your value proposition back to you, sometimes better than you wrote it.

When you do not have PMF, every user describes your product differently. Some say it is a productivity tool. Some say it is a coaching app. Some say it is a Notion competitor. That confusion is the absence of fit.

📌 Use Case: Running a session with a professional services firm that was rolling out a new internal AI tool. I asked ten users to describe the tool in one sentence. I got ten different answers. Six months in, the tool was a solution (hammer) in search of a job (nail). The leadership team thought adoption was the problem. Language was the problem. Adoption followed clarity and alignment.

The AI Move: Use AI to transcribe and analyze customer conversations, support tickets, and reviews. Ask: What phrases repeat across users? When the same phrase surfaces from 5+ users, that is your positioning. Write it on your homepage. Then use it in sales.

Your language checklist:

• Can my users describe my product in one sentence?

• Do their sentences sound similar?

• Am I using their language, or trying to teach them mine?

• What phrase do users repeat that I have not yet claimed?

Stage 5. Watch for Organic Pull

The last PMF signal is the quietest and the most powerful: users bringing you other users, without you asking.

This is not virality. Virality is a mechanic. Organic pull is a behavior. It means your product is solving a problem so specifically, so painfully, that users become evangelists.

Signals of pull:

• Users refer others without a referral program.

• Users write about your product on their own channels.

• Users onboard their friends manually: screen sharing, forwarding, teaching.

• Users get upset if you talk about serving a different segment.

📌 Use Case: With one of the founders I coach, we tracked source-of-signup for 90 days. In month one, 100% of signups came from paid ads. By month three, 40% came from someone who told me about it. That was the moment we knew PMF was real. We shifted budget from paid acquisition to community, and CAC dropped 60% in one quarter.

The AI Move: Use AI to analyze sign-up source data alongside retention data. Are your organic signups retaining better than your paid signups? If yes, your PMF is real, and your best channel is your existing users. If not, you may have marketing-market fit rather than product-market fit.

Your pull checklist:

• What percentage of my new users came from other users last month?

• Are my organic users retaining better than my paid users?

• Am I earning growth, or renting it?

• Who is my product’s biggest evangelist, and have I thanked them?

PMF for Writers and Creators, a Substack Edition

You are not building software. But you are building a product. Your newsletter, course, or community is the product. PMF applies. The signals just look different.

Here is what PMF looks like for a Substack writer:

Signal 1: The Very Disappointed Test (Adapted)

Ask 10 subscribers: “If this newsletter stopped tomorrow, how would you feel?”

• 40%+ say very disappointed or equivalent → You have PMF.

• Most say “I would find another one” → You have interest, not fit.

Signal 2: Open Rates as Retention

• Industry benchmark: 30-40% open rate.

• Your PMF signal: consistently 50%+ open rate.

• Below 25%: Either the wrong audience or the wrong topic.

Signal 3: Reply Rate (Your NPS)

• 1-2% reply rate = normal.

• 5%+ reply rate = strong PMF. Your writing is prompting action.

Signal 4: Paid Conversion

• Free-to-paid conversion benchmark: 5-10%.

• Your PMF signal: 10%+ converting to paid.

• Below 3%: You have readers, not customers.

Signal 5: Organic Subscriber Growth

Are subscribers coming from other subscribers (recommendations, shares, “someone forwarded me this”)?

• 30%+ of new subscribers from organic = real PMF.

• 90% from ads or promotion = paid interest, not pull.

📌 Use Case: A creator I coached had 3,000 free subscribers and 30 paid. She was frustrated. We ran the disappointment test: only 8% said very disappointed. The truth: her newsletter was interesting but not essential. We narrowed the topic from “startups” to “solo founders in year one, month by month.” Six months later: 2,000 subscribers (smaller!) but 250 paid (8x). Fewer readers, but real PMF.

The AI Move: Use AI to cluster your subscriber comments and replies. What topics generate the most replies? What phrases do readers repeat back? That is your PMF signal in language. Write into it.

Writer’s PMF Checklist:

• Would my top 10 subscribers be very disappointed if I stopped?

• Is my open rate above 50%?

• Do subscribers reply to my posts (5%+)?

• Are 10%+ of free subscribers converting to paid?

• Am I growing from readers telling other readers?

Common Writer Traps:

• ❌ Chasing subscriber count instead of subscriber conviction.

• ❌ Writing for the algorithm instead of for your 100 best readers.

• ❌ Adding topics to grow, and losing the people who came for the original one.

• ❌ Confusing viral posts with PMF. A viral post gets attention. PMF keeps it.

The Founder’s AI Discipline (for PMF)

AI can help you see PMF clearly. It can also help you fake it.

• Use AI to measure retention, not to inflate signups.

• Use AI to hear your users, not to guess for them.

• Use AI to run more loops, not to skip the loops.

• Use AI to spot pull, not to manufacture pull.

The founders with real PMF are not the ones with the best dashboards. They are the ones who can explain, in one sentence, exactly why their users would be very disappointed if the product disappeared.

Before You Move to the Next Stage

Ask yourself one question: Am I optimizing for the metrics that would make me proud, or for the ones that would make me right?

Signups feel good. Retention tells the truth. NPS is directional. The Sean Ellis 40% test is the closest thing to a PMF verdict.

The founders I have sat with in their messiest moments (a big launch with no retention, a viral post with no signups, an enterprise pilot with no expansion) did not need more traffic. They needed to see the truth clearly and iterate on the one broken signal.

That is what these frameworks are for.

Bonus: The 5 PMF Signals Scorecard

Before you scale, score yourself honestly.

• Sean Ellis Test: 40%+ very disappointed = ✓

• Day 7 Retention: 15%+ = ✓

• NPS: 40+ = ✓

• Willingness to Pay: 40%+ trial to paid conversion = ✓

• Organic Growth: 30%+ signups from users, not ads = ✓

5/5: You have PMF. Scale aggressively.

3-4/5: Partial fit. Iterate on the weakest signal.

1-2/5: You have interest, not fit. Do not scale.

0/5: Go back to Stress Test 1.

Read more

Build the right thing, faster.

If you are a founder, indie creator, or product person, you already know the problem: it’s never been easier to build something, and never been harder to know if you are building the right thing.

Most teams don’t fail because they can’t ship. They fail because they spend months shipping something nobody asked for.

That’s the problem this Substack exists to solve.

Hi, I am Ruma 👋

I have spent the last decade and a half running design sprints and product discovery work with startups, agencies, and companies, and teaching what I learn in the field at Harvard, Georgetown, and the University of Tokyo. My portfolio includes 3M, Johnson & Johnson, AARP, Tapestry, Walgreens Boots Alliance, Exxon Mobil, midmarket companies, health tech startups, White House Innovation, and local and state governments.

I have facilitated 100+ sprints, helped teams go from a vague idea to a validated prototype in three to five days, and watched firsthand which shortcuts work and which quietly kill products.

That work has taken me into the classroom, too. I have guest lectured and taught executive education at Harvard University, Georgetown University, the University of Tokyo, INTEC Engineering College in Santo Domingo, and the University of Amsterdam. I mentor founders coast to coast in the U.S. and often internationally. In my free time, you will find me exploring innovation hubs in Berlin, Dubai, Riyadh, Abu Dhabi, and Pakistan — with a growing curiosity about East Asia.

I started this Substack because most of what is written about design thinking is either too academic to use or too shallow to matter. A practical approach can not be taught in a one-dimensional manner. There’s a huge gap between empathize, define, ideate as a slide in a deck and actually sitting in a room (or a Zoom call) getting a skeptical team to a testable prototype by Friday.

I write from inside that gap.

I have the receipts to prove it.

What you will get here

Every week, I break down the practical side of building products people actually want:

• Sprint playbooks — step-by-step breakdowns of real sprints: what we tested, what broke, what we would do differently

• Frameworks that fit on a sticky note — design thinking tools stripped down to what you will actually use under deadline pressure

• Validation before code — how to test demand, pricing, and positioning in days, not quarters

• Solo-friendly adaptations — because most sprint advice assumes a 7-person team and a corporate budget. You might be a team of one. I have got you.

No theory dumps. No recycled Medium listicles. Rather, my field notes are from someone who does this work every week.

Strategy execution: the missing piece

Here is a pattern I see constantly: founders and intrapreneurs double down on solving the problem with their idea — sprinting, prototyping, iterating — while the strategy underneath it quietly drifts away from what the market is actually signaling.

Senior leaders reacting to clients miss this piece more times than you can guess.

Strategy gets treated like a one-time deliverable. Calculate your TAM, grab the SAM and SOM numbers, plot them in the three-circles visual, drop it in the pitch deck, done. But those circles describe how big a market is. They say nothing about whether your approach fits where that market is moving — shifting buyer behavior, pricing pressure, distribution changes, competitor moves, timing.

A great solution executed against a stale read of the market still loses.

So, alongside the sprint work, I write about closing that gap:

• Reading market signals and translating them into product and positioning decisions, not just slides

• Pressure-testing your approach against real buyer behavior, not personas on a conference room wall

• Connecting sprints back to strategy so every test moves the business forward, not just the backlog

• The hard calls are knowing when the signals say double down, change the approach, or kill the idea

Validation tells you if the idea works. Strategy execution tells you if the business will.

Free vs. paid — what’s the difference?

Free subscribers get at least one post per week, in the form of essays, teardowns, and ideas worth stealing.

Paid subscribers get the toolkit.

• 🧰 Full sprint templates — my actual design boards, facilitation scripts, and agendas, ready to run

• 📓 Deep-dive case studies — the numbers, the mistakes, and the decisions I can’t share publicly

• 💬 Ask-me-anything threads — bring your product problem, get a facilitator’s answer

• 🗂 The full archive — every playbook and template

Think of it this way: a single sprint facilitator costs $5–15K. A workshop costs $2- 5K. This is $8 a month—less than the cost of sticky notes.

If even one framework here saves you two weeks of building the wrong feature, it’s paid for itself many times over.

Who this is for

• Startup founders who need to validate before they burn runway

• Creators and solopreneurs turning an audience into a product

• Product managers and designers who want to run sprints without waiting for permission

• Anyone tired of just ship it advice that ignores the question of what to ship

• Humans trying out something new, applying to life!

Take it from someone who has not been afraid to start over and try to learn new things. Not because I have to, but rather because I want to.

If this resonates with you, you are in the right place.

Start here

New around here? These three posts will give you a feel for how I think:

1. Getting Stuck? Use Design Thinking.

2. From Stuck to Launched: A Humanistic AI Framework for Getting Started

3. When Busy Looks Like Strategy — and Why the Difference Costs More Than You Think

Building something is hard. Building the right thing doesn’t have to be a guess.

Subscribe for free to follow along. Or go to paid and get the tools I use with clients every week.

Questions? Hit reply — I read everything. Or say hi here or on LinkedIn.

I would love to support you in your journey.

Subscribe now

Updated on Jul 8, 2026. An honest audit revealed a need for a cleaner value proposition, a stronger CTA, and a clearer paid wall value. All posts will be free, with visual artifacts, cheat sheets, and a one-pager for paid subscribers. I am building in public with you.

Updated on Jul 2, 2026. I realized that the section on “Idea Validation for Writers and Creators, a Substack Edition” was missing. It is included now.

Most people do not lack ideas. They lack the moment when someone sat down with them and said: "Here is how you actually start."

I have been in rooms with early-stage founders running out of runway. I have consulted on strategy in complex organizations where nobody agreed on the problem, let alone the solution. I have coached creators who had everything they needed except a clear first step.

The pattern is always the same. Not a lack of talent. Not a bad idea. Just no framework for going from thinking to building.

And now there is a new pattern: founders shipping fast with AI but skipping the thinking. They are confusing speed with progress.

This is my framework. Explained like a human, not a textbook.

A Note on AI

AI has changed three things about how we validate ideas:

1. Building is faster. What took 6 months now takes 6 days with tools like Cursor, Replit, Lovable, and v0.

2. Validating is faster. AI can mine market signals, role-play customers, and analyze interviews at scale.

3. The cost of skipping validation just went up. When everyone can build, the differentiator is who built the right thing.

I weave AI moves throughout this framework as accelerators. Use them when they help you learn faster. Skip them when they substitute for real customer conversations.

Today we start with Idea Validation.

We will cover Product Market Fit (PMF), Go-to-Market (GTM), Pricing, Unit Economics, and Market Research.

Summary visuals for an easy-to-follow framework approach are included.

The Framework: 5 Stages from Idea to Launch

Understand the 4 tests and learn what to measure. Talk to 10 Customers → Problem Validation, Show prototype → Customer Willingness, Assess findings → Solution Viability, Analyze competition → Defensibility.

Use the Interactive Worksheet to make decisions and refer back when pivoting.

Apply the design thinking questions to pressure-test your thinking, see whether you are on track, and identify what may have fallen through the cracks.

Use AI as a thinking partner, not a replacement for customer truth.

At the end of this post, you will have answered these key questions:

• Could you build this if you had to? Do you have the skills or the AI tools to close the gap?

• Would 3+ real customers pay for this? Not hope. Real data.

• What would change your mind about this idea?

• Is this something you would work on for 2-3 years?

TL:DR: Test problems before solutions. Get to yes from 3+ real customers before you commit. In the AI economy, building is cheap. Validating is the new moat.

↳ Next Step: Pick one idea. Run it through the 4 stress tests. If it fails badly, pivot or kill it. If it passes, you have permission to build.

Stage 1. Name the Real Problem

Before you build anything, understand what you are actually solving.

Most founders skip the most important step: understanding the problem before building the solution. Design thinking is not an art exercise. It is how you avoid wasting 6 months on the wrong thing, or 6 days, in the AI era.

They fall in love with their solution before they have truly understood the problem. The quality of your solution is only as good as the quality of your problem definition. A founder in love with his idea asked me this question at the Harvard Health Innovations System global hackathon.

📌 Use Case: Inside companies, this looks like leaders holding on tightly to one product or service. At AARP, I demystified it by creating KPIs with defined frameworks. They were aligned with executive leadership strategies and 3-year goals.

The AI Move: Before you talk to 10 customers, use AI to scan Reddit, Trustpilot, App Store reviews, and forum threads for your problem space. What complaints surface repeatedly? What language do real people use? AI can pre-validate the problem signal, but it cannot replace the conversations. Use it to sharpen your interview questions, not to skip the interviews.

Ask yourself:

• Who has this problem right now, today?

• How are they currently solving it, and why is that not enough?

• What would their life look like if this problem disappeared?

• Am I solving the problem I have or the problem they feel?

Stage 2. Test Before You Build

An MVP is not a cheap version of your product. It is a question with a budget. AI has made this exponentially cheap and fast.

The fastest way to waste six months is to build the whole thing before anyone tells you they want it. Now it is the fastest way to waste six days. The founders who get this ship faster, learn faster, and waste less.

Every podcast, every founder will tell you this. Launch unfinished. If you wait for perfection, you are too late.

The smartest founders I have worked with do not pitch, they test. They find the smallest possible version of their idea that can generate a real reaction from a real person.

📌 Use Case: A young founder I met last year was targeting matchmakers. I took his calls because he was young and hungry. In the first call, he showed me why my hunch was right. He had interviewed top matchmakers across categories and dove into his findings.

The AI Move: Use AI to role-play 5 to 10 target personas before your real interviews. Pressure test your problem statement against these synthetic customers. Catch the obvious gaps in your thinking before you waste a real interview slot. Then go talk to humans.

Your MVP checklist:

• What is the one core assumption my idea depends on?

• What is the cheapest, fastest way to test that assumption?

• Can I get 5 real users to respond to this in the next 7 days?

• Am I building to learn, or building to feel busy?

• Am I using AI to accelerate learning or to feel productive?

Strategy is not a deck. It is the hard choices nobody wants to make until they have to. Choosing what NOT to build is your first strategic decision. AI gives you the ability to build everything. Do not.

👉 Subscribe to get the frameworks I use with founders.

Stage 3. Build the First Real Thing

Done beats perfect at the prototype stage. Every time.

Startups do not fail from a lack of ideas. They fail from a lack of structure, clarity, and someone who has seen the pattern before. That is what this stage is about: bringing in the structure that turns an idea into something real.

With AI tools, a founder or creator can go from idea to working prototype in an afternoon, describing what they want, reacting, refining, redirecting. The barrier to prototyping has collapsed. The question is who acts on that first, and who acts on it thoughtfully.

📌 Use Case: We saw an emerging pattern of adjacent insights during a caregiving experiment. The caregivers were not interested in platforms. They wanted logistical solutions to plan trips for their elderly parent or relative when they could not take them. We had a view into the provider side: missed doctor’s appointments. Empty time slots can decrease a physician’s daily productivity by 12.5% to 14%. For smaller practices, this was estimated at between $50,000 and $150,000 per year.

The AI Move: Tools like Replit, Lovable, Bolt, v0, and Cursor let you ship a clickable prototype in hours. But here is the trap: the easier it gets to build, the easier it gets to build the wrong thing beautifully. Use AI to compress the cycle from idea to prototype to customer reaction. Not from idea to prototype to launch.

Your prototype checklist:

• Does this solve the problem I defined in Stage 1?

• Can someone use it without me having to explain it to them?

• Is it good enough to get a real reaction, not a polite one?

• Have I shipped it to at least one person who is not a friend?

• Did I build to learn, or did I build because AI made it easy?

Stage 4. Read the Room

The skill is not expertise. It is knowing how to move people forward anyway.

I have worked in rooms where nobody agreed on the problem, let alone the solution. Government bodies, early startups, global teams. What I learned: the skill is not having all the answers. It is knowing how to move people forward anyway.

Growing up between cultures teaches you something no MBA does. How to read a room quickly, adapt without losing yourself, and find opportunities others miss. That has been my edge in every consulting room I have walked into. It is the same edge you need here. AI can analyze sentiment. It cannot read a room.

📌 Use Case: I was running a multi-day design thinking session with a leading professional services company. They were rolling out SAP and needed to define their business use cases. Notably, the teams kept bringing the conversation back to change management. The use cases would not work without addressing the elephant in the room. Not part of the agenda, I paused and spent 30 minutes building an argument with the C-Suite direct reports. We incorporated this dependency and quantified the risk.

The AI Move: Record your interviews (with permission). Use Otter, Fireflies, or Claude to transcribe and analyze. AI surfaces what was said. You interpret what was meant. Look for: words used most often, emotional intensity, what got skipped over. The signal lives in what people do not say.

Your feedback checklist:

• What did people do, not just say?

• What question kept coming up that I had not thought to answer?

• What did people ignore, and why does that matter?

• What do I now know that changes my next move?

Growing up between worlds gives you a superpower most people do not even know they are missing: the ability to find the signal in rooms full of noise. In the AI era, signal is the only thing that matters.

Stage 5 — Launch with a Throughline

Transformation isn’t a rebrand. It’s changing how people think, decide, and act.

The founders and creators who launch well aren’t the ones with the best product on day one. They are the ones who can tell a clear, human story about what they are building and why it matters, consistently, under pressure. That’s the hardest thing to do. I have spent years learning how.

Harvard sharpened my thinking. The field tested it. The best ideas I have ever had came from sitting with founders in their messiest moments, not in a classroom. Your launch is your messiest moment. Own it.

📌 Use Case: With founders and intrapreneurs, it's going back to the basics of storytelling. Founders love their baby and want to spend time on every feature. Edit out details and focus on the customer’s problem at hand for budget approvals and VC funding.

Your launch checklist:

• Can I explain what I do in one sentence a stranger would understand?

• Do I know exactly who I am talking to — and what I want them to do?

• Is my story specific enough to be believable?

• Am I launching to the world, or to the right ten people first?

Lead with human-centered principles. Explain like a human, not a manual or an AI bot.

The 4 Stress Tests — Applied to Your Substack

Test 1: Problem Validation
Does the audience care?

• Post 3 pieces publicly before you launch (LinkedIn, Notes, Instagram)

• Measure: Do strangers engage with recognition, not just politeness?

• Pass: 3+ replies from people you do not know saying this is exactly what I needed

• Fail: only friends liked it

Test 2: Solution Viability
Can you sustain the output?

• Can you write weekly for 6 months without burning out?

• Do you have 20+ post ideas already written down?

• Fail signal: you struggled to list 10. Wrong topic or wrong voice.

Test 3: Customer Willingness
Will they actually pay?

• Offer a founding member tier before you hit 100 subscribers

• Pass: 5+ paid signups before launch

• Fail: zero paid interest. You have readers, not customers.

Test 4: Defensibility
Why you, not the next writer?

• Can you state your unique angle in one sentence?

• Weak: “I write about productivity.” Strong: “I write about leading through acquisition from inside the room.”

• Your lived experience is your moat

  ---

👉 Where Great Content Ideas Come From

Source 1: Your Reader’s Language

• Read reply emails word for word

• The phrase that repeats is your next post

• Ask: “What do you wish someone had explained to you?”

Source 2: Your Genuine Frustrations

• What do you rant about to friends for 30 minutes?

• What conventional wisdom in your field do you think is wrong?

• That conviction is a post. And a series.

Source 3: Your Lived Experience

• What have you built, led, or survived that others have not?

• What did you learn that took years?

• Specificity is your moat. Not breadth.

The AI Move (Per Source)

• Reader language: paste replies into Claude, ask “what repeats?”

• Frustrations: describe the rant, ask Claude for the counterintuitive angle

• Experience: describe your career in 5 sentences, ask what here can no one else claim?

Scoring

• 4/4 → Launch it

• 2 - 3/4 → Refine the weakest test first

• 0 - 1/4 → Find a different angle before committing

Common Writer Traps

• “I will figure out the topic as I go” → you will not

• “My friends love the idea” → friends are not your audience

• “I write about everything” → you write for no one (I wrote about 5 topics! and am focusing on one now)

• The best Substack niches sound too narrow at first. That is the point.

One crisp point of view beats ten broad ones. Every time.

The Founder’s AI Discipline

AI gives you superpowers. Discipline determines if you use them well.

• Use AI to validate faster, not to skip validation.

• Use AI to build cheaper, not to build the wrong thing beautifully.

• Use AI to think wider, not to think for you.

• Use AI to draft, not to define your voice.

The founders winning in the AI era are not the ones using the most AI. They are the ones who know when to use AI and when to put it down and talk to a human.

The One-Page Summary

Before You Move to the Next Stage

Ask yourself one question: Am I moving forward because I have learned something, or because I am uncomfortable staying still?

Momentum is valuable. Clarity is more valuable. AI makes momentum cheap. Clarity is still hard-won.

The founders I have sat with in their messiest moments, 90 days of runway, the wrong market, undefined ICP, a great product, did not need more hustle. They needed someone to help them see the pattern clearly and make the hard choice.

That is what these frameworks are for.

Bonus: 4 Stress Tests

Before you build anything, test these four areas.

Then apply the design thinking questions to pressure-test your thinking, see if you are on track, and identify what you may have overlooked.

• Problem Validation: AI pre-screens, humans confirm.

• Customer Willingness: AI cannot fake a $1 commitment from a real customer.

• Solution Viability: AI tools reshape your feasibility AND viability math.

• Defensibility: In the AI era, your moat is data, workflow, speed, or trust, not features.

Read more

Design thinking has become a phrase people reach for when they want to sound innovative. It wins rooms and woos clients — and occasionally it unlocks collaborations that wouldn’t have happened otherwise. But repetition has worn it down. Somewhere, the actual meaning got lost. What it was always really about is simple:

At its core, design thinking is a tool.

It is a good conversation starter for problems that have gotten boring. It won’t save the world, but it can make an old challenge feel worth poking at again.

A genuinely fun way to look at old problems from a new angle. It’s not magic, and it won’t fix everything — but it can make challenges where you keep getting stuck, worth tackling again.

The method itself isn’t complicated.

What it needs to work is simple: everyone in the room has to actually show up. Curious, open, willing to experiment, willing to look a little silly, and willing to be wrong. Without that, no framework saves you.

What’s harder is getting everyone to actually play along — to stay open, embrace a little chaos, try things that might not work, and not flinch when they don’t. This is where behaviors supported by learning and psychology come in, to make sense of momentary pain ahead of a bigger win.

My best sessions have been multi-day boot camps and training with start-ups and students. They are brave, curious, unburdened by the conformity rules of the business world, and ask questions that can push the group to the art of possibility. Participant mindsets and attitudes are as important as the framework itself.

The mind is like a parachute. It works best when open.

Curiosity, almost childlike curiosity, is the secret to unlocking the mindset needed for players to explore. Participants tap into early childlike playfulness to experiment, learn what worked and what didn’t, alter the experiment, test a different route than originally planned, and have fun along the way. Most of all, players must be willing to risk letting go of their million-dollar idea, their ego, sometimes their status, and to come up with new, different, and sometimes crazy ideas they will work on.

The route of applying design can be uncomfortable, and the end is often unknown; nevertheless, like the makings of a good mystery, the fun is in following the trail (business challenge), the players (users), throwing curveballs (barriers), and with a cast of mixed actors (stakeholders with different expertise and unique views of the challenge). This requires switching to the Sherlock Holmes detective mindset.

Political players who are focused on personality rather than the business challenges, unless they are willing to check egos and politics at the door, they will be the least willing participants and can derail energy and dialogue in the room.

The participants who are unwilling to give up the mic are the party poopers bringing everyone down. Gently remind these players what mindsets are encouraged, and introduce them at kickoff. This permits all participants to check one another and provides the language for the right behavior.

Design thinking tactics require even the most seasoned coaches to learn and unlearn, study innovation trends, and be willing to switch techniques in real time based on their audience and the environment. The worst experience I can have in the room is having players with a fixed mindset who are unwilling to be curious. I have paused the conversation to address an organizational challenge that participants might get stuck on that falls outside the realm of design thinking. This is necessary at times. Your willingness to listen will gain confidence and trust.

When do you apply it?

Is there a problem with a user (customer), and you want to understand their persona, habits, and interactions with a product or service?

Maybe you are tackling a challenge that has not yet been solved, or needs a new perspective?

A solution that needs to move to the digital world?

An early idea that needs fleshing out?

A concept that requires testing?

An idea that needs a low-fidelity prototype?

Have dissatisfied customers unhappy with online engagement or product features?

If you have answered yes to any of the questions, pick up human-centered design and apply it to the stage you are in: understanding the challenge, learning who the user is, experimenting with early solutions, prototyping, and learning from the process.

To study a user, whether an individual or a group, is to be a curious detective observing habits, behaviors, and patterns, and, at times, to become an anthropologist studying the evolution of the problem and behavioral trends.

Design Thinking in the Age of AI

AI tools are a gift to the curious mind. They collapse the distance between idea and artifact, letting designers explore more territory in less time than ever before.

Vibe coding is remarkable for ideation, experimentation, and the ability to bring concepts to life. Describe a dashboard, a flow, a gesture, and within minutes, you have something tangible to react to. That speed fundamentally changes how teams can think together.

There is still a case for analog and low fidelity.

Not everything needs to be a polished prototype or a vibecoded to perfection. The rough sketch, the sticky note, the whiteboard diagram — these signal incompleteness on purpose. They invite collaboration rather than critique, and they keep the conversation open exactly when it needs to stay open.

The risk with AI-assisted design is premature closure. When a tool can generate a high-fidelity mockup in seconds, the temptation is to treat it as a decision rather than a provocation. Teams, stop asking, “What problem are we solving?” and start debating button colors.

Design thinking has always been about holding ambiguity long enough to find the right question. AI doesn’t change that imperative — it just raises the stakes. The faster you can build, the more disciplined you need to be about when to build, and what the act of building is actually for.

More on AI integration in future posts.

How do I learn Design Thinking?

Since I published this piece back in 2018, Design Thinking expertise has had its moment and remains relevant. A quick search will reveal the number of courses and experts available. Every educational institution offers courses ranging from a few days to a few weeks, as well as degree programs.

Maybe you are an expert. Even as an experienced coach, I do a refresher now and then.

Courses are offered on a rolling basis. Your learning curve will be richer if you organize a team of colleagues or friends to do it with you. IDEO University, the OG, offers courses online. SAP Learning (previously OPEN SAP) offers introductions to developing software using design thinking, democratizing our quest to humanize online and digital interactions. Stanford d.school has resources.

My expertise, in all honesty, came from leading innovation in multicultural markets at AARP, where design thinking became the platform for innovation. The tools and behaviors were approachable and malleable to turn the legacy organization on its head.

Whichever book, training, or resource you pick up, put it into practice right away. Be willing to toddle along before you learn to stand.

In the comments, please share the challenges you faced in applying design thinking, as well as any success stories. I would love to hear from you.

Additional Resources

• Business Model Canvas, Strategyzer

• Value Proposition Canvas, Strategyzer’

• What is Design Thinking? Why Is It Important HBS (2018)

  ---

This article, “Why Design Thinking Matters,” argues that effective design thinking requires a mindset of curiosity and empathy, rather than merely treating it as a popular business buzzword. The piece highlights that the approach should be used as a structured tool for problem-solving, rather than as a “cool factor,” to address genuine business challenges. It has been updated to reflect wide adoption and best practices.

Variation of this post first appeared on the Mentors Fund blog in November 2018, titled “Why Design Thinking Matters”. Read the article at LinkedIn Pulse.

Learn more about the origin and direction of my Substack

As with most of Vietnam, it's fascinatingly calm on the senses, bar city traffic, and a mix of the old and the new. I particularly enjoyed the strong culture that was noticeably different from north to south in tone and narrative, in colonial loyalties to communist ideals. Its coffee culture is bar none.

Here’s my visual love letter to the ancient city of Hội An, and its remarkable women. A city I often think about, and hands down, the most memorable trip.

The women are front an dcentre of the Vietnamese economy. They are on the bikes, in the shops, running. stands and smiling.

Walking through the market, the women are noticeably the sellers and hawkers. They don’t push you. A lot of things are “one dollar only”. Perhaps a small sum for us, and a decent sale for them. I ended up buying a hand-carved wooden chopsticks case laden with ivory, chopsticks, and fish-shaped ivory chopstick stands.

Further Research & Links -

• Anthony Bourdain in Vietnam: The Complete Country Guide

• Hoi An Travel

• Hoi An, Vietnam (Trip Advisor)

• Hoi An Ancient Town (UNESCO)

Learn more about the origin and direction of my Substack

On June 10th, a room full of nerds gathered at the University Club of San Francisco for HKS on the Road: San Francisco event featuring Jeremy Weinstein, Dean of Faculty, Harvard Kennedy School, and Don K. Price Professor of Public Policy.

The topics teased were eroding trust in institutions, academic challenges, threats to democracy, data centres, wealth inequality, the structural affordability gap, the cost-of-living crisis, massive job losses, economic dislocation, and the notion of democracy as a technology. Weinstein also shared HKS 2036 vision for the future of Harvard Kennedy School, the path to the second century and outlined delivering four imperatives for the future.

Featured HKS alumni speakers were Jeffrey Bleich, MPP 1986, General Counsel, Anthropic, and Rafael Mandelman, MPP 1999, President, San Francisco Board of Supervisors. The 150 alumni in attendance were from classes ranging from the 1960s to this year’s graduates.

Societal Impact of AI, Social Upheaval & The Need for Bold Policies.

The distinguished speakers discuss the societal impact of AI, emphasising the need for collaboration between government, corporations, and the public to address potential risks and dislocations. They highlight the importance of informed democracy and proactive policymaking to navigate rapid technological advancements.

A key theme is the comparison to the Second Industrial Revolution, noting historical parallels in social upheaval, economic change, and the need for bold policy responses such as antitrust laws and an income tax. The speakers advocate for similar decisive action in the current AI era.

The conversation touched on the importance of education, particularly in fostering critical thinking skills and human empathy, to prepare individuals for an AI-driven world. They also discuss the potential of AI to democratize access to resources and services, exemplified by its application in legal aid.

The speakers stress the role of institutions like the Kennedy School in training individuals with public-service orientations and critical-thinking skills, emphasising the need for active participation across sectors to address the challenges posed by AI and maintain democratic values.

Bleich returned to the need for critical thinking at the centre of AI adoption in schools and to avoid relying on ads, a jab at Anthropic’s competitor. He also discussed Dario Amodei’s latest paper, Policy on the AI Exponential and the need for AI regulation.

Some memorable quotes:

• My biggest concern is not that a well-informed democracy can’t solve. It is that we are not a well-informed democracy.

• We have devalued caring for each other.

• We should train to be critical thinkers and good humans

• (today we have to be) Boss of AI Agents

HKS Alumni

Networking at these events is strong.

We relive the deep connections to the school that inspired us to pursue paths in public service, academia, corporate life, and startups. Momentarily, we are transported back to our nerdy student days and connections to the school.

I had a great time chatting with Tushar, working on robotics, Patrick at Google, working on infrastructure, and a former VC turned startup operator, with my friends in leadership roles at Apple, TikTok, and xAI.

As a former HKS Global Alumni Board Director, I saw firsthand the need for our diverse global alumni body to stay connected, learn and upskill, and maintain their bonds. During my term, I established the Digital Innovation Award, chaired two committees, presented Alumni Awards, and highlighted alumni voices in business and startup sectors. During covid, I started the HKS In Tech virtual group with a fellow alum.

I hope that Weinstein will pay heed to this gap in the global alumni body and the desire to connect with minds that have not intersected before, as he continues to introduce new initiatives. This gap persists among the four centres at the school, which focus on their own threads and initiatives without crossing over. A Kennedy School problem, if there was one. I do appreciate the noticeable change in conversation under Weinstein.

A shout-out to Karen Bonadio, Hannah Nguyen, and the entire Alumni Office team for their hard work and commitment to keep the HKS body strong. It’s always a pleasure to see them.

I leave you with a reflection from my 10-year reunion: Ask not what 10 years can do for you.

Share

Note: These global events bring together HKS alumni for an evening of learning and networking to bridge the distance from the Kennedy School.

HKS on the Road is a series of events taking place in cities around the world, giving HKS alumni and friends the chance to meet and hear from our new dean, Jeremy Weinstein. If you have questions, please contact HKS Alumni Events. (source: HKS Alumni Office).

I hope this piece resonated and, if you belong to the alumni community, that this brought back fond memories for you, as it did for me.

Reach out. Tell me what you thought. I would love to connect with HKS alumni who have recently landed in the Bay Area and who are in tech, AI, and startups doing transformational work. Reach out here or on LinkedIn.

Learn more about the origin and direction of my Substack

To receive new posts and support my work, consider becoming a free or paid subscriber.

Subscribe now

Earlier, I wrote about my return trip to Riyadh after twenty years and my impressions of how much had changed. This time around, I came back experienced and armed with more knowledge, and I had a clear list of what I wanted to see during a short trip.

Boulevard: Jetlag and Steps

Fresh off the plane and fighting jetlag, I headed straight to Boulevard. The logic was simple: keep moving, stay awake, and clock some steps while the body adjusts to the time difference. Boulevard delivers on all counts — wide walkways, interesting installations, and enough activity to keep my brain engaged while my body protested the hour. It’s become one of those spaces that works whether you are there for entertainment, food, or just the atmosphere. Jetlagged wandering there beats staring at a hotel ceiling.

JAX District and SAMOCA

JAX District was high on my list, particularly the SAMOCA Museum of Contemporary Art. I had heard about it, especially with a connection there, but I did not make it on previous trips. The museum is set up as an immersive experience across two sections, and you can move through it within an hour, unless certain works pull you in and hold you there. That’s entirely possible. Some installations demand more than a glance. The square footage is smaller than the installations at various MoMas and iconic locations. The experience is unique and eye-opening, showcasing local talent delivering concept creative works that will make you think. The details are incredible.

What made the visit particularly valuable was having my friend Arwa give an insider tour of the space. She walked me through live projects, pointed out which cafes were worth stopping at, explained upcoming developments, and helped me prioritize what to see given my limited time. The district has an energy to it - studios, galleries, creative spaces all concentrated in one area. I left wishing I had more time to explore it properly.

Next trip.

Diriyah: A Heritage Site Worth Revisiting

Diriyah District remains a favorite, and it’s earned its UNESCO World Heritage Site designation. This was the seat of the first government in Hijaz, and ongoing restoration has brought the area back to life, respecting its historical significance while making it accessible to visitors.

Walking through Diriyah, you get a sense of what the original settlement must have felt like - the mud-brick architecture, the narrow pathways, the layout of the old town. The restoration work is extensive and ongoing, but even in its current state, it’s impressive. There’s a particular quality to being in a place that holds that much history, especially when the preservation effort is done thoughtfully.

It’s also less crowded than some newer attractions, making it easier to take your time and actually absorb what you’re seeing. If you have any interest in the region’s history, Diriyah should be non-negotiable on your itinerary. Plan to stay for dinner at one of the restaurants on-site. Expect US prices (with, dare I say, better food). A portion of your ticket purchase goes towards the dining bill.

AlUla: Clearing Up the Confusion

AlUla has evolved into a major destination for entertainment, wellness, and cultural tourism. But there’s still considerable confusion about geography, even among people who should know better.

Let me clarify: Madain Saleh (Hegra) is located within the larger AlUla province. One does not equal the other. Madain Saleh is an ancient Nabataean archaeological site, spectacular, historically significant, and worth a visit. AlUla province is a broader region that includes Madain Saleh and encompasses the old town, contemporary art installations, luxury resorts, adventure tourism, and an expanding calendar of festivals and events.

When people conflate the two, they miss the scope of what’s being developed there. Yes, Madain Saleh is the crown jewel from a heritage perspective, but the province has become a destination in its own right. You could spend several days there and not see everything, ancient tombs one day, rock climbing the next, a concert in the desert that evening.

The scale of investment and development in AlUla is significant, and it shows. Infrastructure is solid, hospitality options range from high-end resorts to more accessible stays, and the variety of activities means travellers of all kinds will find something that works for them.

Here are my detailed posts on AlUla, catalogued under the Travel & Gastronomy section.

The Value of Return Trips

Coming back to Riyadh with more context made a difference. Having family and friends to visit is a big attraction. I had a better idea of what I wished to prioritize, who to reach out to for recommendations, and how to make limited time count. The city keeps adding layers — new districts, cultural spaces, restored heritage sites, and each return visit reveals something that wasn’t there before. It is heavily investing in global cultural events and institutions.

If you are planning a trip, do your research, reach out to people who know the city, and build in flexibility. Things move quickly here. What’s under construction today might be open by the time you arrive.

If you are planning a trip to the region, you may have questions about culture, the length of the visit, or which sites to prioritize.

Reach out. Tell me what’s top of mind. Sometimes, a culturally relevant conversation with someone you can comfortably ask questions to can lead to a memorable trip.

Share

Backing Founders Shaping the Future

Kernel Camp is built on two ideas.

​First, the Middle East and North Africa region produces founders who can compete globally. And second, that world-class founders need to spend time building here in Silicon Valley.

I am seeing this connection firsthand as a way to bridge regions. First, with the scaleups program that I launched with SAP in the US, which aimed to connect the EU, MENA, and the Americas. Second, the ongoing Harvard HSIL hackathon is focused on sourcing and mentoring health startups globally. I am a Venture Advisor at the latter.

​The thoughtful events leading up to the half-day showcase and closing out demo day with the inaugural cohort were impressive.

These startups were selected from the top 3% of applicants from MENA, demoing what they built, shipped, and sold over 8 weeks.

​The founders are some of the sharpest people worth knowing who landed in SF only eight weeks ago, listened to the market signals, pivoted rapidly, and doubled down. They are working on robotics, cybersecurity, AI costs, activation layer for AI agents, and an intelligent operations platform for dynamic AI-guided workflows.

If you would like to connect with the founders and encourage anyone interested in joining the next cohort to sign up, click here.

Live pitches from the Kernel Camp startups 🚀

Also in the room was Mona Ahmad (Reevo), moderating a panel discussion on tech talent from the MENA region in SF-based startups, with Ahmed Rashad (Perle AI) and Ahmad Saeddedin (Corgea YC S23). I was hanging by the edge of the seats. Both founders stressed hiring talent willing to learn, buckle down, and in it to win it.

Sujude Dalieh (Replit) led a conversation with Waseem Alshikh, CTO and Co-Founder of Writer. This was one panel not to be missed. Later, Waseem sat down with the founders for an off-the-record chat.

The event continued with the cohort, guests, and the broader Kernel Camp and Propeller community socializing. We lingered on even after having to vacate the space. I had the opportunity to connect with the teams 1:1 to learn from their experiences and next steps.

I can also confirm that the founders are multifaceted and can extend their skills to different arenas.

​Propeller invests in Seed & Pre-A founders building horizontal AI infrastructure and AI-native applications, from silicon-adjacent to enterprise workflows.

I work with a small number of leaders at a time on focused 1:1 engagements. Not cookie-cutter frameworks or off-the-shelf playbooks. Just clear thinking, honest challenge, and a process built around where you are trying to go.

Reach out. Tell me what’s on your mind. Sometimes, the most valuable thing is simply having someone in your corner who will ask the question no one else is asking.

Learn more about the origin and direction of my Substack

Updated: Jun 8, 2026.

Today, I am announcing a brand new addition to my Substack publication: The Fluid Identity with Ruma subscriber chat.

This is a conversation space exclusively for subscribers — like a group chat or live hangout. I’ll post questions and updates, and you can jump into the discussion.

Here, the subscribers will get a first look at upcoming coaching offers and templates I will publish.

What can we work on together?

How to get started on Substack, Developing your Voice, Startup Pitching, Design thinking templates, to name a few that I am working on.

This will be a space to come and ask for help, get promoted, and brainstorm something that you are stuck on.

Join chat

How to get started

1. Get the Substack app by clicking this link or the button below. New chat threads won’t be sent via email, so turn on push notifications so you don’t miss conversations as they happen. You can also access chat on the web.

Get app

2. Open the app and tap the Chat icon. It looks like two bubbles in the bottom bar, and you’ll see a row for my chat inside.

3. That’s it! Jump into my thread to say hi, and if you have any issues, check out Substack’s FAQ.

Future Proofing at Mentors Summit 2026

Every year we meet to take stock: of our portfolio, our thesis, our conviction. But this year feels different. This year, the ground itself has shifted.

Not gradually — but the way that earthquakes shift things. Fast, irreversible, and impossible to ignore.

We are living through the greatest technological inflection since the Internet.

Artificial intelligence is no longer a feature, a sector, or a trend. It is the operating system of the next economy. Every industry — healthcare, logistics, finance, education, media, manufacturing — is being rewritten from first principles.

The companies being founded right now will become the defining institutions of the next thirty years. And the companies that fail to adapt? Some of the largest names you know today won’t survive the decade.

The Mentors Fund Summit 2026, therefore, centered on an urgent theme: Future Proof. Not because we believe the future is certain — but because we believe the right entrepreneurs, with the right mindset, can build companies that thrive in any version of it.

Throughout the summit, our partners, entrepreneurs, and guests explored how we can safeguard our work and our lives in a world increasingly defined by AI. Yet, as we moved through firesides, panels, and pitches, I realized our gathering was future-proof in a much more fundamental way.

Share

Moving with Conviction

So what does it actually mean to be future-proof in the AI age?

It doesn’t mean predicting the future with precision. No one can. It means building in a way that doesn’t require you to be right about everything — and positions you to win regardless of which path the technology takes.

We think about it through the five pillars. Every company and every entrepreneur should be asking themselves the following questions —

1. AI-Augmented, Not AI-Dependent - Are you using AI to amplify human judgment — or building fragility by removing it entirely?

2. Proprietary Data Moats - When models commoditize, unique behavioral, relational, or operational data becomes the last durable advantage.

3. Adaptive Business Models - Your revenue model must evolve faster than the technology disrupts it. Build for iteration, not just launch.

4. Human-Centered Trust - As AI scales, trust becomes scarce. Companies that earn and protect it command pricing power no algorithm can replicate.

5. Speed as Core Discipline - The half-life of competitive advantage is shrinking. Learning and pivoting faster than rivals is now a survival skill.

The pace of change is relentless. The capital environment is demanding. The technology is advancing faster than any roadmap can accommodate.

But here’s what history tells us: the most important companies in the world were built during disruption, not despite it. Amazon was built in the rubble of the dot-com crash. Airbnb launched into a financial crisis. The greatest venture returns in history were seeded in chaos.

During the Summit, the audience heard from portfolio companies demonstrating exactly what future-proofing looks like in practice, from us on where we are placing our conviction, and had real-time opportunities to connect with the people in the room navigating similar terrain.

Mentors Fund is a community of funders and founders, yes — but more than that, we are a community with a shared bet on the future. A bet that great businesses, built with integrity and vision, still win in the long run.

Join us at the Mentors Summit 2027 next May.

Subscribe now

Venture capital investment in AI has reached unprecedented levels, with 2025 serving as a year of massive growth and Q1 2026 shattering all previous records in a single quarter.

The primary trend is extreme capital concentration, in which a handful of “foundation model” companies (such as OpenAI and Anthropic) are absorbing the majority of global venture capital dollars.

AI VC Funding Overview (2025–2026)

Total AI VC Funding in 2025, $211B – $258.7B

AI captured 61% of all global venture capital (OECD). In Q1 2026, a single quarter exceeded the entire funding total of 2025 (PitchBook), estimated at $255.5B.

2025: The Year of Dominance

According to the OECD and Crunchbase, AI became the undisputed engine of the venture market in 2025.

• Market Share: AI startups accounted for 61% of global VC investment value ($258.7 billion out of $427.1 billion).

• The “Mega Deal” Era: Roughly 73% of all AI investment value came from deals exceeding $100 million.

• Geographic Lead: The United States remained the primary hub, attracting roughly 75% ($194 billion) of global AI deal value.

• Top 2025 Deals: Massive rounds were closed by Anthropic ($13B), xAI ($10B), and Scale AI.

2024 vs. 2025 vs. 2026 (Q1)

The growth has been exponential, driven by the massive capital requirements for compute and infrastructure.

2024: The Foundation

AI funding reached approximately $114B, nearly doubling from the previous year as generative AI went mainstream.

2025: The Infrastructure Surge

Funding jumped to ~$254B. Investors shifted focus toward “compute-heavy” startups, with infrastructure and hosting alone attracting $109B (OECD).

2026: The Record-Breaker

In just three months of Q1 2026, AI startups raised $255.5B. This surge was driven by “Horizontal Platforms” (OpenAI, Anthropic, xAI), which accounted for $197B of the total.

2026: The Winner-Take-Most Market

As of June 2026, the market is characterized by massive, multi-billion-dollar “super-rounds” for established leaders.

Top AI Funding Rounds (Q1 2026)

1. OpenAI: Closed a record-breaking $122B round in March 2026, valuing the company at over $850B.

2. Anthropic: Raised $30.6B in February 2026.

3. xAI: Secured $20B before being acquired by SpaceX in a landmark $250B deal.

4. Waymo: Raised $16B (Series D) to lead the autonomous machines segment.

Emerging Sector Shifts

While horizontal frontier models (LLMs) take the most cash, specialized sectors are gaining traction:

• Defense AI: Reached an all-time record of $14.6B in early 2026, nearly doubling its 2025 total of $9.6B.

• Autonomous Machines: Generated $29B in Q1 2026, more than triple the value of the previous quarter.

• Vertical AI: Interestingly, while funding amounts are high, the number of deals in vertical applications (AI for specific industries like Law or Health) fell to its lowest level since 2018, signaling that investors are becoming much more selective.

Note on Sources: Data compiled from 2026 reports by PitchBook, OECD, Crunchbase, and KPMG Venture Pulse. Funding totals can vary slightly between providers based on their definition of “AI-related” companies.

I work with a small number of leaders at a time on focused 1:1 engagements. Not cookie-cutter frameworks or off-the-shelf playbooks. Just clear thinking, honest challenge, and a process built around where you are trying to go.

Reach out. Tell me what’s on your mind. Sometimes, the most valuable thing is simply having someone in your corner who will ask the question no one else is asking.

Learn more about the origin and direction of my Substack

I had a difficult work situation because I decided to research, dig deep, ask questions, and start the initiative, the old way. The human brain needed time to absorb the new content, reach my own conclusions, and form the perspective that was asked of me.

It was too slow, I was told.

That gave me pause.

The expectation was AI slop at AI speed, without first doing any foundational diagnostic work, the opposite of change management 101. Act as a consultant, without the history and context, and produce standalone solutions. Why would this not work in my experience? Whether as an in-house expert or an external consultant, success came from learning the context - business model, challenges, culture, historical decisions, the motivation for change, and so on.

Here are some real conversations happening in my chats right now.

A good listen on why Continual Learning matters by a16z, covering in-context learning, reasoning, and retrieving from context frozen in time without updates from real experience. I like the framing that explains context and continual learning to leaders who don't grasp how AI models are trained and deployed. Baseline - in-context continual learning is critical, with humans in the loop

Telltale signs

To predict a situation such as the one I landed in, ask these questions:

1. Does leadership have a clear vision, and can articulate it?

2. Does leadership lay out the strategic plan in tangible terms and the steps that need to be taken?

3. Are leadership and team strategic, or do they manage projects?

4. What's the history, portfolio of work, and the tolerance for change?

5. What work has been delivered that is quantifiable and tracks impact?

6. What are the culture and communication trends versus what you are being told?

7. Does the team have psychological safety, and is there visible evidence?

8. What's the understanding and experience with AI?

Here's the brass tacks: many teams are inexperienced with AI tools and are jumping on the bandwagon without a reasonable amount of time to learn and do the work. It’s a combination of pressure from the top, not having the time, and not knowing where to start,

Maybe I should have cheated and produced AI slop. After all, the evidence for organizational tolerance was apparent.

Additional reading ICYMI -

I work with a small number of leaders at a time in focused 1:1 engagements. Not cookie-cutter frameworks or off-the-shelf playbooks. Just clear thinking, honest challenge, and a process built around where you are trying to go.

Reach out. Tell me what’s on your mind. Sometimes, the most valuable thing is simply having someone in your corner who will ask the question no one else is asking.

It felt like the right moment to revisit the key concepts from Leadership On The Line by Ronald Heifetz and Marty Linsky, in which Technical Problems with clear-cut solutions are prioritized over the more arduous and complex Adaptive Challenges that require systemic change.

The framework has been taught for over four decades at the Harvard Kennedy School of Government, with executive education and commissioned workshops. It is probably the toughest course you will ever take, requiring serious reflection and a disruptive method of teaching, by design.

Core Concepts of Adaptive Leadership

“Going Forth to Die” — The Nature of Leadership

• Leadership is inherently dangerous. It requires challenging the status quo and disrupting comfortable norms

• Those who lead change often become targets — historically and organizationally, leaders are “sacrificed” by the group when they push too hard

• The act of leadership is therefore an act of courage — knowing the risk and proceeding anyway

• The Authoritative vs. Authoritarian Distinction — Adaptive leaders use their authority to direct attention to the work, not to provide all the answers

Technical vs. Adaptive Problems — Technical problems have known solutions. Adaptive challenges require people to change their values, beliefs, or behaviors. Most leaders mistakenly apply technical fixes to adaptive problems

Loss — At the heart of all adaptive challenges is loss — people resist change not because they fear the future but because they fear what they must give up

Disequilibrium

• Change and growth require a period of productive disequilibrium — a deliberate destabilization of existing norms, habits, and comfort zones

• Too little disequilibrium = no change; too much = chaos and system breakdown

• The leader’s job is to regulate the heat, keeping it high enough to motivate change but not so high that the system collapses

• This is called managing the “Productive Zone of Disequilibrium.”

Ripeness — Some adaptive challenges can only be addressed when the group is “ripe” — when the pain of the status quo exceeds the fear of change

Factions

• Any organization or community contains competing factions, each with its own values, loyalties, and definitions of the problem

• Factions will often fight each other rather than confront the real adaptive challenge

• Leaders must identify factions, understand what each faction stands to lose, and broker across them

  

• Work Avoidance — Groups will consistently find ways to avoid the pain of adaptive work through scapegoating, denial, or displacement onto side issues

The Holding Environment — A safe enough container (trust, relationships, structures) that allows a group to tolerate the stress of adaptive work without flying apart

Loyalties

• People in authority have multiple, often conflicting loyalties — to constituents, to the organization, to their own values

• Adaptive leadership requires leaders to sometimes disappoint their own base or faction to serve the broader mission

• Loyalties can be weaponized by factions to neutralize a leader (”you have betrayed us”)

• Giving the Work Back — The leader’s role is to resist the temptation to solve problems for people, and instead return the adaptive challenge to those who must do the work

Assassination — Literal and Organizational

• Groups under stress will attempt to silence or remove leaders who threaten their equilibrium

• “Assassination” in organizational life takes many forms: marginalization, character attacks, distraction, seduction, or outright removal

• This is not personal — it is a systemic defense mechanism of the group resisting change

Staying Alive — The Survival Imperative

• “Getting off the dance floor and onto the balcony” — leaders must develop the ability to zoom out and observe the system they are part of, without losing presence within it

• Build a personal support network outside the organization that isn’t dependent on your role

• Distinguish between your role and your self — attacks on the role are not attacks on your person

• Pace yourself — adaptive work is a marathon, not a sprint

  ---

If any of this resonates with you, you have been, or are in a position where you have to lead the difficult work — I would like to have that conversation.

I work with a small number of leaders at a time on focused 1:1 engagements. Not cookie-cutter frameworks or off-the-shelf playbooks. Just clear thinking, honest challenge, and a process built around where you are trying to go.

Reach out. Tell me what’s on your mind. Sometimes, the most valuable thing is simply having someone in your corner who will ask the question no one else is asking.

Please look beyond the clichés manufactured by the hate industry, funded with nefarious intent, to propagate hate and demonize my community.

Enjoy the many ways we celebrate togetherness and faith.

This post is best viewed on mobile.

Eid-ul-Adha 2026 began on Wednesday, May 27, 2026, and lasts three to four days, depending on the region. Also known as the Festival of Sacrifice, it commemorates the Prophet Abraham’s willingness to sacrifice his son, Ishmael (Ismail), in obedience to God.

The holiday is marked by communal morning prayers (Salat al-Eid), acts of charity, festive family gatherings, and the Qurbani (animal sacrifice). Source: ISNA.

In North America, it is not yet recognized as a public holiday. San Francisco leaders Assemblymember Matt Haney and Supervisor Bilal Mahmood are spearheading a legislative push for Assembly Bill 2017 to officially designate both Eid al-Adha and Eid al-Fitr as recognized state holidays in California.

In the news today:

• Meanwhile, Boston’s Mayor Wu extends Eid greetings.

• Mamdani, NYC’s first Muslim mayor, joins New Yorkers in prayer for Eid Al-Adha and see the Mayor’s instagram post.

Eid al-Adha is a major driver of food security.

Globally, an estimated 50 million animals are sacrificed during Eid al-Adha each year. This translates to roughly 2 billion to 2.5 billion pounds of meat processed in just three days, with each animal yielding about 40 to 100+ pounds of consumable meat depending on the species. In countries with large Muslim populations, millions of animals are processed in a matter of days. For example, Pakistan accounts for 6 to 9 million animals, while Bangladesh sees over 10 million. Source: Swistas Consultancy.

Meat Consumption & Distribution Rules
The traditions of Qurbani ensure widespread consumption by dividing the meat into three distinct portions:

1. One-third is retained by the family performing the sacrifice.

2. One-third is shared with friends and relatives.

3. One-third is donated directly to the poor and disadvantaged.

   Source: Wikipedia.

Share

Eid Mubarak!

This morning, I went to the Eid prayers, where I would see friends in the Bay Area, ate breakfast, and had chai from local restaurants and vendors. I volunteered for cleanup, a recurring theme in my service to the community, at celebrations run by well-meaning people. I met new people and realized serendipitous connections between us! Eid is communal, it is inclusive, and it removes barriers, making it accessible to Muslims, especially those who are alone, traveling, or away from their families. I called my family, and with the time zone differences, it will take me 3 days to reach everyone!

Subscribe now

We are living through a loneliness epidemic. Depression and suicide rates are climbing. People go days — sometimes weeks — without a real conversation with another human being. Connection, once the most natural thing in the world, has become something we have to consciously seek out.

At a meditation talk during the pandemic, I asked teacher Tuere Sala how to navigate the political upheaval gripping this country — the sustained attacks on Black and brown communities, the ambient dread. She reached for the words of Thich Nhat Hanh: sangha will save us. Community will save us. The message was to be with each other, to rely on one another.

Human beings were not built to wake up alone, commute in silence, make someone else rich for eight hours, and return home to a curated screen. We were not designed for that kind of life — and the data reflects it. The U.S. Surgeon General had formally declared loneliness a public health epidemic. And this isn’t just an elderly problem. In a hyper-digital age, the crisis is reaching preteens and teenagers, young people whose sense of self is still forming. That is a dangerous and urgent trend.

Communal spaces, shared norms, tight-knit communities, and public social contracts are not relics of the past. They are medicine.

Meaningful social connections are fundamental to mental health, a recurring theme in blue zone research, and — as evidence continues to mount — a powerful buffer against the neurological disorders that so often surface in later life.

Share

An Architecture of Life

Islamic tradition has long understood this. The concept of ummah — the global community of Muslims bound not by blood or geography but by shared values and mutual responsibility — places collective wellbeing at the center of spiritual life. You are not just accountable to God. You are accountable to each other. That accountability is lived out daily, in small and large ways. From our most sacred rituals to the ordinary rhythms of daily life, our faith is rooted in the call to seek genuine connection and do good for one another.

In Muslim households, extended family living is not a burden — it is the architecture of life. Grandparents are not outsourced to facilities; they sit at the dinner table. Children grow up watching their elders be honored, which teaches them, without a single lecture, what honor looks like. Young people have daily access to multiple generations of wisdom, humor, grief, and resilience. The home becomes a living classroom. Hasan Minhaj called out this lack of respect and honor for elders in the West as a major problem, one that threatens politics and power.

The tradition of Jumu’ah — the Friday congregational prayer — is another example. It is not optional to socialize. It is a weekly, communal obligation that pulls people out of their private lives and into shared physical space. You stand shoulder to shoulder with your neighbor, your colleague, a stranger. You pray together, eat together, and check on one another. The mosque functions less like a church in the Western sense and more like a community center with a spiritual heart.

This has been coming up a lot lately, and it resonates: In order to live in a village, you have to be a villager. Translation - you have responsibilities and duties as a villager to strengthen the community bonds, help build, contribute, and safeguard the community in good faith, words, and action. In other words, be a good human being.

My friend Tarek shared his conversation with on people who go to the mosque regularly yet still lack social connections. Note: this is not a religious conversation, it centers on the human need to be seen and have a social connection. Even in these communal spaces, social etiquette can be lacking, with a tendency toward the dry ways of the administration or caretakers, understandably focused on operations and day-to-day running. Interestingly enough, they mention having “greeters,” like at Walmart, and imams who reach out to folks in the community to check in, greet them, and ask if they need support. The stories are endearing.

Then there is sadaqah — voluntary charity — and the culture of hospitality so embedded in Muslim life that turning away a guest is almost unthinkable. You feed people. You show up. You do not let someone grieve alone or celebrate alone.

These are not just religious practices. They are technologies of connection, refined over centuries, designed to keep people from falling through the cracks.

The Western world has largely abandoned these structures without replacing them. We have traded the village for the algorithm. And we are paying the price.

So here is the call. Not a grand one. Not a movement or a hashtag or a pledge.

Just this: say hello.

To the person next to you on the train, staring at the same window. To the stranger at the café nursing their coffee. To the brother or sister standing beside you at Jumu’ah, whom you have seen every Friday for two years and never learned their name. To the person next to you in yoga class. Say hello. Introduce yourself. Mean it. When I do that, people are surprised and will respond with warmth.

Don’t expect a soulmate. Don’t expect a deep conversation. Expect only what it is — a civil human exchange, a smile, a moment of mutual recognition that says I see you, you are not invisible, neither am I. That is enough. That is more than enough, because most people are walking around starving for exactly that.

Normalize apologizing. When you snap at someone, when a misunderstanding festers, when frustration leaks out sideways — name it. I was short with you, I am sorry. That came out wrong. I was having a hard day. These are not signs of weakness. They are the stitches that hold communities together.

Community is not built in grand gestures. It is built in ten-second interactions, repeated daily, by people who decide that the stranger in front of them is worth a moment of their attention.

Be that person. Start today. Start now.

A note on Dhull Hijjah (pronounced Zull Hij’jah).

Hate towards many communities, including my Muslim community, is rising. With the devastating incident at the Islamic Center of San Diego, scholars and experts have highlighted the rising bigotry and hate speech from politicians, media and influencers using disgusting language and tropes.

In America, this is unacceptable.

I share below in the spirit of understanding and outreach, dating back to the father of Abrahamic traditions, Prophet Abraham (Ibrahim) & his son Ishmael (Ismail). May we all do better and accept our differences in faith and practice.

Why is it a Sacred Time?

Dhul Hijjah is highly revered for several key reasons:

• The First Ten Days: The first ten days of this month are considered the most virtuous days of the entire year. According to Islamic tradition, good deeds and acts of worship performed during this brief window are exceptionally beloved to Allah.

• The Hajj Pilgrimage: The “Month of Pilgrimage” is the only time Muslims can perform Hajj in Mecca, Saudi Arabia. It is one of the Five Pillars of Islam, a once-in-a-lifetime obligation for those who are physically and financially able.

• The Day of Arafah: Taking place on the 9th day of the month, this is the spiritual peak of the Hajj. Millions of pilgrims gather on the plains of Mount Arafat to pray and seek forgiveness. Muslims worldwide who are not on the pilgrimage often choose to fast on this day, which is believed to expiate the sins of the past and the coming year.

• Eid al-Adha: Occurring on the 10th day, this is the second major Islamic holiday. It commemorates the absolute faith and devotion of Prophet Abraham (Ibrahim) and his willingness to sacrifice his son, Ishmael, upon Allah’s command. To honor this, Muslims who can afford it perform a sacrifice (Qurbani), sharing the meat with family, friends, and the less fortunate.

Courtesy: Islamic Relief Worldwide.

I have tried hot yoga, CorePower, Bikram, Hatha, Vinyasa, modern-day variations, and I have gone with the flow since my teens.

Before I found my way back to the mat, I dabbled with other forms of movement. I gave pilates a real shot. I built a gym routine I was proud of and showed results. I was doing the things you are supposed to do, checking the boxes, showing up. But something was missing — that particular feeling of being both challenged and held at once.

Then I found a studio within walking distance of home. That detail matters more than it sounds. There’s something about removing the friction — no commute, no excuses, just shoes on and out the door — that makes consistency possible. The last time I had a fabulous yoga studio within walking distance was fourteen years ago in DC. The studio offered two community classes taught by yoga instructors, and the instructors collected their teaching hours. They were welcoming, warm, and generous. I went once out of curiosity. Then again. Then I couldn’t stop.

What followed was a slow and deeply satisfying experience in trying different instructors. Each one brought their own rhythm, their own emphasis, their own way of cueing a pose. Some were meditative and slow. Others pushed you. Some sessions are hot, others are not. I am listening to my body and trying the non-heated classes. I have been returning, started noticing, and kept learning how my body was responding.

Subscribe now

The agility of a twenty and even thirty-year-old body was gone. It’s precisely why I am sticking with the poses that used to be easy. Now sajdas are more satisfying, rukuhs feel like a good stretch, salaams headturns are good neck twists, and getting off the ground feels like a power move.

Yesterday's beach yoga was anything but serene — 20mph winds, sand stinging our eyes and mouths, bodies swaying on shifting ground. Frustration crept in quickly, until our teacher's calm voice cut through it all with a quiet reminder: life rarely unfolds the way we plan.

The changes came gradually — which, I have learned, is how the best changes come. Strength arrived first, quietly. Then stability, that underrated quality that makes everything else in life feel a little more grounded. I am still working on it. And then one day, I caught a glimpse of my arms in the mirror. My upper arms, usually the first thing to soften when I step back from weight training, looked different. More defined. Strong.

My gym routine is on pause. I assumed that I would pay a price for that. So I did what any mildly anxious person does: I went online researching the trade-offs. Was I losing muscle? Missing something essential that only weightlifting could give me? Was I compromising my fitness?

The answers were more reassuring than I expected.

Yoga, it turns out, is not a consolation prize. It builds functional strength, improves mobility, challenges your cardiovascular system, and — perhaps most importantly — keeps you returning to it. Consistency beats intensity almost every time, and yoga has cracked the code on keeping me consistent. Almost every day is a good habit.

I am not arguing that yoga replaces everything. It doesn’t, and I will be back in the gym soon, most likely with better body awareness than I had before. But right now, in this season, the mat is where I belong. It’s where my body responds, rebuilds, and surprises me with a quiet strength.

Some returns take longer than others.

This one was worth the wait.