Way back when I was working for Wolters Kluwer in APAC, the defence for why you’d continue a subscription to their legal service was “no lawyer can go into a court with a print out from google as their precedent” True but it didn’t mean you couldn’t get the same insights for cheaper. This is the crisis impacting consulting today - the myriad of lazy Partners who’ve help to grab a headline by not reviewing the grad or juniors AI generated slop hasn’t help their credibility.

The lazy version says consulting is about to get obliterated because AI can now write, summarise, synthesise, and generate respectable looking deliverables in seconds. The defensive version says nothing important changes because clients still buy trust, judgement, and political cover. Both stories miss the awkward middle.

What interests me is that AI may not kill consulting cleanly, it may make midmarket consulting weird. That matters because a lot of midmarket advisory work has always rested on an unstable mix of real expertise and expensive packaging. Not fake work, exactly, but not always deep magic either. Plenty of firms have been paid well to take a messy situation, impose a structure on it, create a persuasive narrative, and hand back a cleaner looking plan. That can be useful, in many cases it is useful, but it also means the category has always contained a layer of value that depends on the client not seeing too clearly how much of the offer is pattern recognition, formatting, and controlled ambiguity.

AI starts messing with that layer first, not the hardest part, not the boardroom judgement call, not the delicate stakeholder management, not the operator who has seen this exact kind of failures three times before and knows where the traps are hiding. Those things still matter, in some cases they may matter more, but a surprising amount of the midmarket consulting middle is exposed once decent synthesis becomes cheap and fast.

If a client can now get a competent first pass on market framing, operating hypotheses, workshop outputs, slide structure, interview synthesis, and scenario packaging without paying consulting rates, something changes. The question is no longer just whether the firm can produce a polished answer, the question becomes what part of the answer actually required them.

That is where the market starts to feel awkward.

The strongest firms will move upwards, they will lean harder into judgement, domain depth, hard client specific calls, implementation credibility, and senior operator involvement. They will make the human part more explicit, not less, they will stop pretending the value is in document production and start showing that the value is in knowing what to do when the document is wrong, incomplete, politically inconvenient, or impossible to execute.

The weaker firms have a harder problem, they may discover that a lot of what looked like expertise was really well packaged pattern recognition - Good slides, confident phrasing structured summaries, a bit of industry language and a plausible operating model. Enough ambiguity to sound sophisticated without being sharply accountable, AI is already uncomfortably good at that zone.

That does not mean clients will suddenly stop buying external help most clients are not looking for a machine to dump words on them, they are looking for someone to absorb uncertainty, create momentum, align stakeholders, and help make decisions that feel expensive or risky. Trust still matters, accountability still matters, being in the room still matters, but the economics around those things will need to change.

The old consulting bundle becomes harder to defend when some of the most visible outputs can be produced much more cheaply. It gets harder to charge premium rates for generic synthesis when the client knows generic synthesis is abundant. It gets harder to hide behind process theatre when a first pass strategy deck no longer signals unusual effort and it gets harder to keep large teams attached to work that increasingly looks like it could have been compressed.

That is why I do not think the clean story is “AI replaces consultants” the sharper story is that AI forces a rebundling. Some work gets compressed, some gets exposed, some becomes more valuable because the fluff around it gets stripped away. A client may become less willing to pay for the production of an options paper but more willing to pay for the person who can look at three plausible options and explain which one will fail inside their organisation and why, that is a different kind of premium.

It is less about presentational polish and more about difficult judgement under messy conditions. This is where midmarket consulting gets weird, because the middle has historically depended on blending those two things together. Enough real insight to sound credible, enough packaging to feel premium, enough ambiguity to preserve room for manoeuvre. AI weakens that blend, it makes the packaging cheaper, the generic parts become easier to imitate. Best of all it raises the pressure on firms to prove that what remains is not just nicer wording around familiar patterns. the ability to respond to this pressure will not land evenly.

Boutiques with real specialist depth may do well, firms with genuine operator credibility may do well. Advisory shops that can combine AI enabled speed with sharp judgement may even widen their advantage. But firms that built their model on expensive synthesis without strong implementation credibility may find the centre of the market getting slippery under their feet.

Clients will notice too, they may not say, “this used to take you three weeks and now I know a machine can do half of it in an afternoon” but they will feel it. They will ask harder questions about scope, they will become less impressed by the polished artefact itself. They will want to know what thinking happened beyond the visible output. The customer will become more alert to whether the team in front of them is actually making hard calls or simply arranging information elegantly. This is healthy conflict.

It may force the category to become more honest about where value really lives. Consulting is not about to disappear. But the old middle may not stay comfortable. AI makes it easier for clients to separate real judgement from expensive synthesis, and once that separation becomes clearer, some firms will move up, some will get compressed, and some will discover that what looked like expertise was more fragile than they thought.

That is why I think AI is about to make midmarket consulting weird, not dead or unsafe just much harder to bluff in.

I think one of the most interesting things about AI right now is that the real advantage is starting to shift. For a while, progress looked like sheer activity, more prompts, more retries. more model swapsm, more chains, more everything. That made sense when the tools were still new and people were still discovering what they could do, but that is not where the next layer of value comes from.

The strongest AI users are often the ones who know when to stop, that might sound less exciting than the usual AI theatre - but i’m not here to be your AI hyper - but I think it is actually a more useful and optimistic idea. Once AI becomes normal inside everyday work, the advantage stops coming from enthusiasm alone, it starts coming from judgement.

Knowing when a rough draft is already good enough is judgement, knowing when a task needs a deeper pass is judgement, knowing when the next token is likely to improve the work versus merely soothe your own uncertainty is judgement. And knowing when the job should leave the model entirely and go back to a human is judgement too.

That is why I think one of the most useful habits in AI is effort discipline, most teams still do not talk about this clearly. instead they talk about prompt quality, model quality, tool choice, access, governance and more recently cost. All of that matters, it does, but a lot of waste appears one layer lower, in the moment by moment behaviour of users who have not decided what level of effort a task actually deserves.

A quick internal summary might need one competent pass and a fast human skim, a sales contract for $100m in product probably needs a slower model, a source check, and a final human rewrite. A messy strategic question may not belong with AI at all until somebody has first made the thinking sharper. Treating those jobs as if they all deserve the same depth of model effort is not sophistication. It is just lack of discrimination, and this is exactly where I a lot of AI overuse begins.

Some users reach for another prompt because the first answer was weak, fair enough. But some reach for another prompt because they have no stopping rule. They keep polishing a low stakes task until the total effort has outgrown the value of the work itself. Others bounce between models not because the job genuinely needs a different capability, but because it feels active and intelligent to keep trying something. In practice that often means paying more, waiting longer, and still not making a clearer decision.

The more mature move is often simpler, decide what “good enough” looks like before you start. As Stephen Covey wrote in 7 Habits of Highly Effective People “Start with the end in mind” this has never been more true than now. Don’t prompt for a task, prompt for a finished thing, then describe everything that thing must be true to.

If the job is low risk, repetitive, and easy to verify, use the cheapest level of effort that reliably clears the bar. If the work is high stakes or publicly exposed, spend more attention where it counts. If the problem is still poorly formed, stop pretending that another model pass will rescue a vague brief. Fix the brief or bring in a human.

This is why I think restraint is becoming a status signal. For a while, advanced AI usage looked like visible cleverness. The person who knew all the tricks seemed ahead, but the next wave of serious users may look calmer than that. They will know which tasks deserve frontier horsepower and which ones do not, they know when to escalate and when to skip AI entirely. Experience will teach them when another pass is real improvement and when it is just expensive hesitation dressed up as diligence.

That matters commercially as well, as firms push AI into more ordinary work, bad effort discipline will get surprisingly expensive. Not only in tokens, although that matters, also in time, delay, and managerial noise. Low value loops slow decisions down, endless retries create the illusion of rigour while quietly making simple work take longer. Teams start paying twice: once in model usage and again in human attention spent reviewing and reworking things that should have been settled earlier.

The goal isn’t to use less AI for the sake of frugality, it is to match effort to stakes.

That distinction matters, heavy AI usage can be perfectly sensible when the value is real, nobody should be proud of saving tokens on work that genuinely deserves deeper analysis. The point is not minimalism, the point is discipline, mature users do not confuse more model activity with better judgement.

I suspect this will become one of the cleanest divides between amateur and professional AI use. Amateurs will often keep going because stopping feels risky. Professionals will often stop because they know the job has already crossed the quality bar. Amateurs will keep spending to avoid making a call, professionals will make the call.

That is why I think one of the next real AI advantages is not better prompting, it is knowing when to stop.

I think one of the most interesting shifts in AI right now is that good judgement is starting to become a product feature. For the past year or two, a lot of the conversation has lived at the wrong level. Vendors talked about model intelligence, buyers talked about seats, access, and pricing, finance teams talked about spend controls, operators talked about prompts. All of that matters, but it leaves out the question that increasingly matters most inside a real business: does the system help people use AI well?

That is why I think Microsoft Copilot Cowork’s /cost skill is more interesting than it first appears.

Not because this is really a story about Microsoft praise. And not because the market desperately needs one more article about AI pricing. What interests me is the design principle hiding underneath it. A simple skill that makes the cost of effort visible at the point of use can do a lot to improve behaviour.

That matters because I have already been circling this theme from a few directions. In AI Cost Is Becoming a Behaviour Problem, the core point was that the bill is not the real issue by itself. The behaviour creating the bill is the issue. In The Smartest AI Users Know When to Stop, the point was that maturity starts to show up in effort discipline rather than endless visible activity. In The Next AI Advantage for ANZ SMEs Is Managing Work Better, the argument was that supervision, routing, thresholds, and review are becoming management capabilities in their own right.

This feels like the next step on from all three, the world has moved from abstract norms into product patterns. In other words, we are starting to see tools embody the behaviours that earlier pieces could only argue for in theory.

That is what makes /cost interesting, it gives the user a clearer sense of what a task is actually worth. It reminds them that the goal is not to minimise spend at all times, the goal is to spend properly, those are very different ideas.

If a Cowork task is genuinely strategic, messy, high-leverage, or likely to move the dial, then spending $15 on it can be perfectly rational. In fact, being too cheap on important work can be its own kind of stupidity. If a better pass produces a clearer strategy, a stronger decision, or a more valuable outcome, then the spend is doing real work.

But the inverse is just as important. If someone spends even $2 on a task that ordinary Copilot, a lightweight assistant, or a standard chatbot could have handled just fine, that is not good discipline. That is not serious adoption. It is just poor judgement wearing a thin costume of modernity.

That is where Satya Nadella’s line about using frontier models for frontier problems becomes genuinely useful. It is a simple management rule disguised as model strategy. Not every task deserves the deepest model, the richest tool, or the most expensive path. Some tasks deserve premium effort. Many do not. The skill is knowing the difference, and the commercial opportunity is designing systems that make that difference easier to see.

That is why I think the most useful lesson here is broader than Microsoft, simple skills may turn out to be more important than they look.

The market still gets distracted by the biggest demos, the most autonomous agent stories, and the loudest claims about replacing whole functions. But a modest skill that clarifies cost, effort, routing, or escalation inside the workflow can sometimes be more commercially important than another round of abstract capability theatre. It helps ordinary users act with better proportion. It teaches better instincts. It nudges people toward the right level of effort before waste, delay, and confusion start compounding.That is a powerful thing!

A lot of AI governance still assumes behaviour will be fixed afterwards. Finance will notice the bill. IT will notice the sprawl. Leadership will notice the confusion. Then somebody will draft a policy. The better path is to shape the behaviour upstream. Make good judgement easier. Make overuse more obvious. Make escalation feel deliberate rather than default.

This is why I do not think the future belongs only to the firms with the smartest models. It may also belong to the firms with the clearest effort signals. The organisations that handle AI well will not just ask whether aa model can do the task. They will ask whether this task deserves this level of spend, this level of reasoning, and this level of attention. They will build operating habits around that question. And increasingly, the best tools will help them do it.

That is the larger opportunity hiding inside something as simple as /cost. It is an example of a broader category of product design that treats judgement as part of the workflow, not as a lecture delivered afterwards. It shows how a small skill can reinforce better stopping rules, better routing, and better economic discipline without forcing every user to become a mini procurement committee.

That is where I think the world has moved forward, we are no longer just arguing that AI usage needs better norms, no, we are starting to see the first useful examples of software that can teach those norms in practice.

And if that pattern spreads, the winners may not just be the people with access to more AI. They may be the people whose tools quietly help them use it with better judgement.

In part one I wrote about the first five challenges and the friction they expose, the next five put pressure on it. This is where leadership can stop being politely curious about AI and start asking whether managers are still protecting weak systems with personal effort.

Challenge #6: “Ban one type of work from reaching leadership unless it meets a standard”
Leadership Challenge: From next


This is where you really get going with AI, as a regular reader you will know that the vast majority of my content is free for everyone as I want to help accelerate AI adoption and ensure that NZInc is a leader not a lagger in AI Adoption. You can spend thousands of dollars on a “AI Consultant” or you can subscribe to get part 2.

Read more

Earlier this week? Well the previous article I teased about this in https://everdayai.substack.com/p/ai-is-changing-the-operating-model. Here’s the conversation you as a founder, owner, CEO need to be having with your leadership. Let’s go…

If I were running a SME in Australia or New Zealand this week, I would stop asking for an AI strategy deck and start issuing sharper operating challenges. That is where the practical gains are. Not in another round of tool comparisons, but in forcing managers to show where the business is still running on unwritten rules, slow approvals, vague ownership, and senior people wasting judgement on work that should arrive cleaner.

The challenges are structured in a way that gives you the actionable challange and my rationale behind them. Please take them, tweak them, make them yours and get moving on your companies AI journey today.

Here are the first five.

Challenge #1: “Pick one workflow where approvals are clogging up growth”
Leadership Challenge: By Friday, show me one recurring workflow where too many routine decisions are still climbing back to a manager, and propose one category of case that should no longer need manual review.
Why: This works because it forces a team to expose a real bottleneck rather than talk vaguely about productivity. In a lot of businesses, the growth drag is not lack of effort. It is quotes, credits, campaign sign-offs, or customer exceptions that keep crawling back up the ladder because nobody has turned judgement into a rule.

Challenge #2: “Write down the thresholds your business is still keeping in someone’s head”
Leadership Challenge: List the five commercial or operational thresholds we rely on most, but have never written down properly: discount limits, margin floors, escalation triggers, refund rules, overdue follow-up timing, VIP customer handling.
Why: Most firms are running on invisible thresholds that feel obvious until the one person who knows them is away or overloaded. This challenge surfaces the hidden rules that keep work socially managed instead of operationally managed, which is exactly where a lot of AI and workflow frustration starts.

Challenge #3: “Identify one place your senior people are wasting judgement on cheap work”
Leadership Challenge: Find one place where a senior person is still spending time reviewing, correcting, or re-routing work that should arrive cleaner before it gets to them.
Why: This is one of the best tests in the whole list because it usually reveals a more expensive problem than bad tooling. In an accounting firm, it may be a partner fixing client notes that arrive without supporting documents. In a distributor, it may be an operations lead checking every escalation because the queue does not separate late but manageable from genuinely high risk. Either way, senior judgement is being spent too low in the stack.

Challenge #4: “Audit one queue that nobody really owns”
Leadership Challenge: Pick one queue that creates drag every week, sales follow-up, support tickets, supplier issues, debtor chasing, compliance reminders. Show who owns it, where it stalls, and what could be standardised or handled as a first pass by AI.
Why: Every (yes every) company has at least one queue that belongs to everyone socially and nobody operationally. That is where dropped balls, duplicated effort, and fake busyness love to gather. Make one manager map the queue properly and you usually find a cleaner operating fix within a day or two.

Challenge #5: “Define what good enough actually means for one repeat task”
Leadership Challenge: Choose one repeat task and define what a good output must include before it reaches senior review.
Why: A surprising amount of rework comes from nobody ever making the quality bar explicit. A quote should show assumptions, exclusions, and margin band. A client update should include source reference, owner, and next step. A support reply should make the decision clear. Once that standard exists, human coaching gets easier and safe AI use gets easier too.

None of those five asks a manager to become an AI expert by Friday. All five do something more useful. They expose where the business is still leaning on memory, interruption, and senior rescue instead of better rules. That is usually the difference between firms that get practical value from AI and firms that keep talking about it.

If a leadership team worked through just those five this week, it would already have a much clearer picture of where the real operating leverage sits. Up next, 5 more actionable challenges, stay tuned, subscribe, even support my writing for as little as a coffee a month, it means a lot to me.

Most AI writing still assumes the prize is labour reduction. Get the model, trim the admin, automate the drafting. Push more work through the same payroll base. Fine, sure if that’s what you think success lots like then make it happen. But a lot of ANZ SMEs are not actually constrained by a shortage of first drafts, they are constrained by something more ordinary and more expensive: too much work waiting for a busy human to notice it, judge it, fix it, or pass it on.

That is why I think the next useful AI story for small and mid-sized firms is not really about automation, it is about management leverage. Spoiler I have some interesting challenges for Leadership to issue to Management coming

Put it bluntly businesses do not need more output first, they need a way to scale management quality before they can afford another management layer. That is a sharper claim than the usual line that AI “saves time”. Time savings are nice, but they are not the real strategic shift. The more interesting shift is that AI can start taking some of the load that sits between work being created and work being properly managed.

In a lot of firms, that middle layer is a mess, a founder checks every quote over $15,000 because margins are thin and one bad number can wipe out the profit on the job. A partner rewrites client emails because junior staff are directionally right but still too soft, too vague, or too risky. An operations lead keeps a mental list of the customers who must never miss a promised delivery window, then spends half the day firefighting the queue. None of that shows up on a product demo, it is still where the business lives or dies.

For the last year, the good AI framing was that work was moving from chat to workflow, and I still think that was right. Chat was the training wheels that we all needed, but the money sits in systems that actually move jobs forward.

Now there is a second shift. Once AI is doing more first-pass drafting, triage, follow-up, classification, and exception spotting, the operating model of management starts to change as well. Managers do less constant touching and more rule-setting. Less chasing, more threshold design, less blanket checking, more exception review. Less rewriting the same kind of weak output, more fixing the system that keeps generating it. That is not a small change, but it is what better management looks like.

Take a 25-person accounting firm. The obvious AI use case is drafting memos faster or summarising a tax change. Useful, but not very interesting. The more valuable use case is a system that produces first-pass client notes with source references attached, marks any claim that is not tied back to a document, and automatically routes anything involving payroll, trust structures, or unusual deductions to a senior reviewer. Suddenly the manager is not cleaning up every note by hand. They are reviewing the handful of cases where the risk is real and the judgement is worth paying for.

A distributor with thirty or forty staff has a different mess. Customer service, warehouse updates, and delivery issues all collide in the same day. The operations manager becomes the human router because nobody trusts the queue to sort itself. A better AI setup is not a charming chatbot in the CRM. It is a system that reads incoming issues, drafts the update, tags the likely cause, and escalates only when the order value is above a threshold, the delay is about to breach an SLA, or the customer has already had two misses this month. That is management leverage. The manager gets fewer interruptions, and the interruptions are better.

Trades businesses have their own version. Usually the owner is still acting as the commercial backstop because quoting, supplier drift, and schedule changes all carry margin risk. A decent AI layer could draft the quote from prior jobs, compare material costs against last month’s supplier pricing, and stop cold if the gross margin falls below, say, 18 per cent or the scope falls outside the usual pattern. That last part matters. The owner should not be reviewing every bathroom renovation quote. They should be reviewing the weird extension with custom steelwork and half-finished plans.

The strongest objection here is obvious enough - but dating very quickly and just sounds like a naysayer - Plenty of managers will say AI still makes too many subtle mistakes, so shifting more supervision into a system is dangerous. Fair point, in some workflows it will be dangerous. If the data is messy, the approvals are vague, or the commercial rules live only inside one person’s head, AI will amplify the confusion rather than fix it. This is exactly why the opportunity sits with management first.

The firms that get value will not be the ones pretending the system is magically correct. They will be the ones that finally make their own standards explicit: what needs evidence, what needs approval, what can be cheap and fast, what needs a slower and more careful pass, what always stays human. In other words, AI rewards businesses that learn how to manage work deliberately instead of socially.

That is also why I do not think the real comparison is “manager versus AI”, no the comparison is between two operating models. In the first, the business runs on interruption, memory, and heroics (firefighting anyone?) In the second, the routine work is routed, checked, and surfaced in a more consistent way, and the human judgement gets concentrated where it has the highest value.

Small firms should care about that more than big firms, not less. Large enterprises can afford committee-heavy governance and a lot of expensive theatre around process. SMEs usually cannot. They need something leaner. They need a way to stop every issue, every draft, and every awkward exception from climbing the ladder to the same two or three people.

That is why I think the next AI advantage for businesses is not just more output per employee, it is the chance to build a better management system before the business solves the problem the old-fashioned way and hires another layer of coordinators, checkers, and internal go-betweens.

The practical question is not “where can we use AI?” It is harsher than that.

Why and Where are our best people still wasting judgement on routine supervision?

Find that answer and you are usually close to the real use case. Not the flashy one, the valuable one.

The first big win won’t be a fully autonomous workflow, it may however be a manager who stops spending Tuesday morning clearing a backlog of small approvals, weak drafts, and follow-up gaps, because the system already handled the ordinary cases and surfaced the three that actually deserve attention. That is not a side benefit. For a lot of SMEs, that is the entire ball game.

Stay tuned for how to challenge management and actually get them moving.

Here is the most uncomfortable sentence in our first month of running consumption-based AI. We asked several hundred people to manage a budget they could not see.

The mechanics are simple enough, every task a user delegates consumes credits, each person has a monthly ceiling: one hundred dollars, or roughly seven thousand credits, for most people, and two hundred dollars, or roughly fourteen thousand, for heavier users. Reach the ceiling and the tool stops, with no reset until the first of the following month. What does not exist, is a running balance, no user can open anything and see how much of their month remains.

What they have instead is a slash command (/cost). type it after a task finishes and you learn what that task consumed. It is a receipt, issued after the money is spent, for one transaction at a time. Our official guidance is to run it after every task and use the number to build an instinct for what different kinds of work cost, so that a cheap result becomes a signal to reach for a lighter tool next time.

The remarkable thing is how well it has worked, a receipt at the moment of spending turns out to be a better teacher than a dashboard nobody opens, and people who ran the command a few times changed what they delegated. The problem is a user who most needs the signal is the least likely to ask for it.

So we compensated where we could. Caps enforced at the platform, spend attributed to a named cost centre, and a fortnightly nudge telling each person in plain language where they sit against their allowance. That last one is the actual balance, delivered by email, because the product could not deliver it in the app, Cowork did the heavy lifting here of course.

Lesson: The economics flipped, and the estate is now hybrid

This is not a switch from fixed to consumption, it is a combination of both, layered across the estate, often on the same users and the same tools. Our AI spend runs is forecast and budgeted for an is climbing, spread across licensed Copilot, our enterprise AI platforms, pay-as-you-go consumption, developer tooling, analytics and CRM agents. Cowork is one meter among many, which is why the governance frame has to bind every tool rather than only the newest one. Governance is how we hold both acceleration and predictable cost, in healthy tension.

Lesson: The persona ceilings, and the surprise about executives

We set dollar-denominated ceilings by persona rather than credit allowances, because dollars are what a manager can reason about.

The counter-intuitive finding is that executives are the cheapest persona. Senior leaders mostly review and summarise, which a standard licensed assistant handles perfectly well, so several are candidates to revert to a plain Copilot licence. Writing that in a post aimed at executives is disarming, and it makes the wider point better than any policy statement: the people with the most authority to buy the capability are often the ones who need it least.

Lesson: Agents break the old arithmetic, and every reason is a lever

Six mechanisms explain why agentic spend behaves unlike a licence. Agents resend accumulated context at every step, up to a thousand times the tokens of a single chat. Roughly 60 per cent of an agentic task’s cost is the agent checking and repairing its own work. The same task can cost thirty times more from one run to the next, which means cost is a distribution rather than a price. Extended reasoning on simple work is pure overhead, poorly designed hand-offs between agents multiply spend for no gain in quality, and prompt verbosity moves the bill enough that concise prompting cut some workflows by 30 - 40 per cent.

Each of those is a control we own rather than a cost we absorb: model routing, prompt hygiene, orchestration design and per-run budgets.

Lesson: Model choice is a first-order cost decision that users make unknowingly

We ran an identical prompt across two frontier models. One returned in 68 credits. The other took 524, close to eight times the cost for the same piece of work. Nothing in the interface told the user that. Our guidance is now blunt: stay on automatic selection or the smaller model by default, try the research and analysis agents in Copilot before reaching for anything premium, and reserve frontier models for genuinely frontier problems. Model routing is our highest-leverage cost control and it currently lives entirely in user habit.

Lesson: The silent burn is a scheduled job, not a person

Our mental model of runaway consumption was an individual doing something extravagant. This was not the case! The real event was a set of scheduled dashboards rebuilding hourly in the background, which consumed a full monthly allocation in twenty-four hours with nobody in the room. Scheduled and background work bills exactly like interactive work but generates no human moment of hesitation. Instrument it separately, alert on it separately, and make someone own it by name.

Lesson: The bill and the behaviour live in different systems

This is the plumbing detail that catches everyone, the dollars and the per-user detail sit in two different planes, cloud cost management holds the actual spend and the commitment drawdown, but not per user. The productivity suite’s credit report holds who consumed what, but not the dollars. Per-user recharge is only possible by combining the two: take each user’s credits from the usage report, value them at the credit rate, then reconcile the total back to the dollars on the subscription. The billing platform alone will never tell you who spent what.

Lesson: This is an industry pattern, not our anomaly

The external evidence matters because it stops the story reading as a local failure. Token prices have fallen roughly two hundred and eighty-fold while enterprise spend on large language models tripled in twelve months. Ninety-three per cent of surveyed enterprises are over budget and a fifth have already throttled AI use over operating cost. Around ten per cent of users drive about sixty-five per cent of consumption, which is precisely why graduated caps should target the right people rather than everyone. The line to end on is not ours, but it should be: tokens are not value, tokens are the bill.

This is a two-part series that’ll reflect on the real lived experiences of consumption models in the enterprise. We started using Cowork about 4 months ago on unlimited plans and although we had a sense that it would be a chargeable tool, the realities of the first month on a consumption model are quite different.

The mental model changed, not just the tool

Copilot behaves like an intern and Cowork behaves like a teammate. The framing that landed best in our onboarding sessions was that Cowork is not a reset but an extension of existing AI habits. Copilot does one small task at a time and you check everything it produces. Cowork takes the whole job end to end and brings it back for you to review and approve. Executives understand this immediately, because it is a delegation story rather than a software story.

Ask, assist, build, delegate

The most useful thing we published was a coherence line across the whole estate:

Autonomy, capability and cost all rise from left to right, inside the same trust boundary. It gives people somewhere to stand before they choose a tool, and it reframes cost as a consequence of delegation rather than an arbitrary tax. We paired it with a borrowed line that does a great deal of work in a room of executives from Satya Nadella “you do not need frontier models for everything, you use frontier models to solve frontier problems.”

Reaching for the most powerful tool is the expensive instinct

Our governance position paper predicted this before general availability, users reach for the most capable tool for everything, including work that a standard licensed assistant would handle at no incremental cost. We also knew the vendor’s own guidance on when to use what would not be ready in time, so we published our own decision card first. That turned out to be the highest-return piece of change material we produced, and it lets us say honestly that we saw the behaviour coming.

Skills are where the compounding value showed up

The demonstrations impressed people, the reusable skills changed their week, s fortnightly executive update built from four prior examples went from roughly four hours of work to about thirty minutes, and a first-draft presentation skill did the same job for another team. The pattern that works is to build the skill after a recurring task, by asking the tool to turn a finished output into a repeatable recipe. The value is not the prompt. It is the institutionalised prompt.

We onboarded narrowly, and then went quieter

Access is off by default. Users request it through our service portal, with manager and cost-centre approval, on the same pattern we already used for licensed Copilot. One month in we have gone further rather than opening up, we agreed a soft launch on a pull basis with no broad publicity, and we deliberately chose a topic other than Cowork for our continuous engagement campaigns.

An adoption team choosing not to drive adoption is genuinely counter-cultural, and it is the most quotable decision in the whole programme. The reasoning is simple arithmetic, if everyone holding a Copilot licence (for us that’s measured in thousands) used Cowork, our run costs would compound 10x in a month. Manager approval alone will not hold that line, because managers approve readily until they get the bill.

Segment the base, or over-serve the ready and under-serve the cold

We mapped four populations rather than one: the enabled cohort, the eligible pool of licence holders, the chat users with no licence, and the wider AI-capable base. That mattered because Copilot adoption across our business units ranges from 30 per cent to 99 per cent, so a single go-live pack would have over-served the ready and under-served the cold. The uncomfortable corollary is that the least active users carry the highest bill shock risk, which means readiness nudges and conservative starting caps matter most exactly where engagement is lowest.

We ran the rollout on the tool itself

The nudge campaign that went to our enabled cohort was built by the capability we were governing. A usage spreadsheet went in, users came out banded by consumption, and personalised guidance was drafted and queued for approval: tailored change communications to all Cowork users in ten minutes of human effort. Heavy users received capacity protection advice and light users received encouragement. If your governance process cannot be run by the capability it governs, you probably do not understand the capability yet.

Next i’ll talk about tokenomics and the uncomfortable truths that all organisations are going to have to face over the near turn. This isn’t just a switch it off and ignore it problem, this is not going away. Operating and Business models need to shift.

To put this into context we have a team who estimated a project would take 3 years, they’ve recut that to 12 months. What is delivering a project in 1/3 of the time worth to you, and how much would you reinvest in Tokens? This is what we all need to be able to answer.

The AI labour story is being told too crudely! The usual headline is familiar by now. AI comes in, jobs go out, companies quietly reduce headcount, and the org chart gets thinner. I do think some of that is coming. Pretending otherwise is silly. But I also think that framing skips the part that happens first inside serious organisations.

Heavy AI adoption may create hiring demand before it creates layoffs. That is not because AI is underpowered. It is because useful AI adoption is operationally demanding. Before a company can remove meaningful human load from the business, it usually has to add or empower people who can redesign workflows, wire systems together, define guardrails, track outcomes, and keep the machinery alive once it is in production.

In other words, the first shortage may not be labour in general. It may be the specific kind of labour needed to make AI useful.

This is why I get impatient with the lazy attrition story. On a spreadsheet it sounds clean. If AI lifts productivity, perhaps you simply stop replacing some people and let the organisation shrink over time. That may happen in places. But as a real operating strategy, it is too passive, and it usually ignores the actual bottleneck.

The actual bottleneck is capability. If your firm does not have enough people who can translate messy work into bounded AI-enabled systems, then AI does not magically solve your labour problem. In the short term, it can expose how weak your internal execution capability really is.

That is why I think the first hiring wave will often be more practical than people expect. Not giant research teams, not moonshot projects. More likely workflow engineers, automation builders, integrators, ops-minded developers, and domain people who understand the work well enough to reshape it.

That layer matters because AI does not self-install, somebody has to decide what job the system owns, where the human still reviews, how failures get handled, how the output is measured, and what gets fixed when performance drifts. That is not background noise. In some firms it becomes an entirely new layer of work. This is also why I think the better comparison is not AI versus jobs. It is capability building versus passive cost cutting.

An organisation that leans mainly on attrition may save some money. An organisation that hires or develops a small number of useful builders can create a compounding internal advantage. One good person who can ship ten bounded AI workflows into finance, support, operations, and compliance changes throughput across the company. That is a very different outcome from quietly leaving vacancies open and hoping the remaining team just absorbs the load.

That does not mean layoffs never come. Some roles will shrink. Some coordination layers are going to be under real pressure. Some kinds of admin work will need fewer people over time. But the sequence still matters.

1. First you build the capability.

2. Then you stabilise the workflows.

3. Then you learn where real substitution actually exists.

Most public commentary jumps straight to the third step and treats the first two as if they do not matter.

I also think this changes how leaders should talk to their teams. If executives frame AI only as a cost-saving instrument, they create defensive behaviour. People protect their patch. Managers avoid redesigning work. The company optimises for fear instead of throughput.

If leaders frame AI as a capability-building shift, the discussion gets more honest. Where do we need builders? Which teams are bottlenecked by repetitive coordination work? Which functions can carry more volume if they get better systems? Which roles become force multipliers if we back them properly?

That is the adult version of the conversation.

It is also probably what many smaller firms will actually do. They are not going to wake up and replace half the company with AI. They are going to hire, back, or stretch a small group of people who can build leverage into the business. Later, that may reduce the need for extra coordinators, admins, or managerial padding. But the first move is often additive in capability even if it later becomes subtractive in headcount.

If I were advising a leadership team now, I would ask a few simple questions.

• Which jobs in this business are constrained by repetitive coordination work?

• Who here can turn that work into a bounded AI-enabled workflow?

• Are we assuming labour savings before we have built the capability required to deliver them?

• Are we backing the few people who could multiply everybody else’s output?

Those questions are more useful than another dramatic argument about which profession disappears next.

The labour story is still moving, but I think one thing is already clear. The first real organisational effect of AI may be a scramble for builders, not a clean wave of layoffs. The firms that understand that will make better decisions than the ones still treating AI as either an instant redundancy machine or a shiny toy.

I think a lot of firms are still asking the wrong AI question. They are asking which model to buy, which vendor to standardise on, which assistant has the best interface, or which lab is slightly ahead this month. Those questions are not irrelevant. They are just no longer the main event.

The more useful question now is much blunter: can you actually get AI installed into real work? That is why I think AI has stopped being mainly a tools problem. It is becoming a deployment model problem.

The gap that matters is no longer between firms that have AI licences and firms that do not. It is between firms that can point to real workflows now running faster, cheaper, or with less human drag, and firms that are still mistaking experimentation for progress.

I keep seeing the same pattern. A company buys access to a model. It runs a few workshops. Teams are told to experiment. Somebody shares a prompt pack. Someone else gives a presentation about the future of work. Six weeks later there is a lot of language about momentum, but the actual business process still looks the same. The queue is still manual. The approvals still bottleneck. The handoffs are still clumsy. The AI lives in a chat window instead of inside the work.

That is not really a tools failure. It is a deployment failure.

The most interesting signal in front of me right now is not that models keep improving. They obviously do. The more important signal is the growing attention on the people and systems that can get AI embedded into production workflows. Cursor’s forward deployed engineer story is one example. The broader enterprise chatter points the same way. The hard part is not getting access to intelligence. The hard part is getting intelligence into the workflow without creating a new mess around it.

That means the firms pulling ahead are building a different capability. They are learning how to take a messy business job, narrow the scope, define the boundary, assign an owner, set the review points, track the outcomes, and improve the system once it is live.

That is the deployment model. It sounds less exciting than a frontier-model debate, which is probably why people underrate it. There is no glamour in saying your advantage is workflow mapping, exception handling, audit trails, and boring operational discipline. But in practice that is exactly the machinery that turns AI from an interesting tool into operating leverage.

This is why I think a lot of current model talk is slightly misleading for ordinary firms. Yes, model choice matters. Latency matters. reliability matters. Provider risk matters. Cost matters. But if you cannot reliably ship one useful AI workflow into finance, support, legal ops, or sales operations, then a marginally better model does not rescue you. Your bottleneck is not the model. It is your ability to deploy.

That is also why I would be careful with organisations congratulating themselves for “having an AI strategy” when what they actually have is tool access plus PowerPoint. A real AI strategy now looks more like a rollout system. It knows how to choose the next workflow, launch it with guardrails, measure whether it worked, and repeat the pattern somewhere else without starting from zero every time.

If I were sitting with an ANZ leadership team right now, I would not ask to see another demo. I would ask to see the deployment in motion.

• What workflow are we changing?

• Who owns it?

• Where does the AI act?

• Where does the human review?

• What is the success metric?

• What happens when the system gets it wrong?

• How quickly can we use the same pattern in the next part of the business

  Everyday AI is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Those are the questions that separate AI theatre from real capability.

I also think this changes which people matter most inside the organisation. Early on, the loudest voices were often prompt enthusiasts, innovation leads, or vendors selling possibility. In the next phase, the important people are the workflow owners, integrators, engineers, ops leaders, and practical people who actually know where the process breaks in real life. The centre of gravity is moving from ideas to installation.

That is the shift I think leaders need to absorb. The winners are not just buying intelligence. They are learning how to deploy it repeatedly. And once that becomes a repeatable internal habit, the model itself stops being the whole story.

Here’s the thing I think a lot of people still miss about AI adoption: the next wave probably does not begin with enterprise transformation programs, governance committees, or another round of chatbot trials. It begins when small businesses quietly start using persistent AI agents to run real work in the background.

That matters because small businesses do not buy technology to admire it. They buy tools to survive. If an always-on AI agent can chase unpaid invoices, reconcile transactions, triage support requests, draft campaign variants, prepare month-end packs, or keep scheduling moving without constant prompting, that is not a novelty. That is operating leverage.

For the last two years, most AI discussion has been trapped in the frame of the prompt. Someone asks a question, the model answers, and we all debate whether the answer was good enough. But that framing is already getting old. The more important shift is from AI as a tool you talk to, toward AI as a worker that stays attached to a job.

That is why small business is such an important proving ground. Big companies can afford to spend twelve months in pilot mode. They have process owners, procurement reviews, security gates, change programs, and political reasons to move slowly. A 40-person firm does not have that luxury. If the owner believes a persistent AI system can reduce admin drag in finance, support, or marketing, they will try it quickly because the pain is immediate and the upside is obvious.

This is also where the product market is finally getting more mature. The winning AI products for this segment will not look like generic chatboxes. They will look like job-shaped systems. The promise is not “ask me anything.” The promise is “I keep your receivables moving,” “I prep the monthly close,” “I route and answer first-line support,” or “I keep your campaign engine warm every day.”

That packaging matters. Small businesses do not want to become AI labs. They want outcomes. They do not want to orchestrate five models, tune prompts every week, and build evaluation harnesses for sport. They want something that behaves like a junior operator who never forgets the checklist, never leaves the queue untouched, and keeps improving as the business learns where the bottlenecks actually are.

This is also why I think the labor conversation needs to get more honest. A lot of executives still frame AI as a headcount-reduction story, usually through attrition. The logic sounds sensible on paper: if AI helps, maybe you simply stop replacing some people and let the org shrink naturally. Sometimes that will happen. But as a primary strategy, it is too passive, and in many cases it misses the real opportunity.

A young AI engineer, even an imperfect one, can be more valuable than another round of quiet attrition because they actively create operating systems rather than just reducing salary lines. Attrition is subtraction. A capable early-career AI engineer is multiplication. They can wire tools together, design workflows, build lightweight guardrails, instrument what works, and turn scattered AI experiments into repeatable internal machinery. One person who can reliably ship ten useful agents into finance, support, sales ops, and marketing can create more enterprise value than eliminating two or three roles and hoping the remaining team somehow absorbs the work.

That is the strategic distinction. Replacing headcount through attrition saves cost. Building agentic capability changes throughput. The first helps the P&L a little. The second changes what the company can do.

And there is another uncomfortable truth here: many businesses will discover that the first real return from AI does not come from replacing experienced people. It comes from increasing the output of the people they already trust. A finance lead with an always-on close assistant becomes faster. A support manager with an always-on triage layer handles more volume without drowning. A founder with an AI ops layer can run a cleaner business without hiring an entire middle layer prematurely. In other words, the best early use of AI may be to delay managerial bloat and administrative sprawl, not to hollow out the core team.

That is why the small-business angle is so important. SMBs are where the economics are clearest. If an AI worker saves six hours a week in invoicing, five hours in support triage, and several days a month in reporting and reconciliations, nobody needs a manifesto to understand the value. The result shows up in speed, fewer dropped balls, and less founder exhaustion.

If I were advising a small or mid-sized company today, I would not start with “how do we roll out AI?” I would start with “which recurring jobs are slow, repetitive, expensive, and ugly enough that nobody really wants to own them?” That is where always-on agents belong first.

A few examples of reasons to act now:

• If invoices regularly go out late, build or buy an AI workflow that drafts invoices, checks missing fields, nudges approvals, and follows up on overdue accounts.

• If support inboxes clog up every Monday, deploy an agent that categorises requests, drafts replies, escalates urgent items, and summarises trends by product or customer segment.

• If month-end close is still a fire drill, use an agent to collect source files, flag anomalies, prepare variance notes, and assemble the first reporting pack for human review.

• If marketing output depends on one overworked person, set up an agent to generate campaign options, repurpose content across channels, track performance, and keep the content calendar moving.

• If the founder is the human API for everything, use agents to prepare status summaries, scheduling options, follow-up drafts, and operating dashboards before decisions get blocked.

None of this requires pretending the systems are perfect. It requires choosing bounded jobs where partial autonomy is still valuable. That is the practical threshold most companies should care about. Not “can the AI do everything?” but “can it do enough of this recurring work that my team gets leverage every week?”

That is the shift I think we are about to see. The AI employee probably does not arrive first as a polished digital executive inside a Fortune 500 boardroom. It arrives as a background worker inside a small business, quietly handling the operational sludge that has always consumed too much human time.

And once that starts working, it will spread very fast. Because unlike the last wave of AI adoption, this one will not be judged by demos. It will be judged by whether the work got done.

I’ve been watching the AI conversation settle into two comfortable extremes.

At one end, Copilots a chat window bolted onto existing software, they answer questions, draft emails, summarise meetings. Useful in the same way a better search bar is useful, it augments the person doing the job, but it doesn’t change the job.

At the other end: AGI the vision deck your execs have been sold on. Machines that can do everything a human can do, only faster and cheaper. This is what gets the keynote applause and the billion dollar valuations. It’s also years away, and nobody in your organisation can product plan for it.

Between these two bookends is a large, fertile territory that almost nobody is naming, I’m calling it the agentic middle.

The technology for this is already good enough, the bottleneck isn’t capability. It’s that nobody’s thinking in job-shaped units, we’re still buying tools that assist people when we could be buying tools that execute processes. If you’re an IT leader having the AI conversation with your executive team, this is the frame that lands. Not “we need more GPU” not “we need to hire ML researchers” It’s:

“pick one process that costs you three people and has clear inputs and outputs. Let’s make that process run itself.”

I keep hearing the same AI case studies “we built a chatbot that answers HR questions” or “our sales team uses a copilot to write proposals.” Fine, awesome you build a RAG (Retrieval Augmented Generation) pipeline, these are the appetisers, the main course looks different, and for a SME isn’t that much more complicated.

Lets take accounts payable reconciliation. A supplier sends an invoice, then the agent reads it, matches it against the purchase order, flags discrepancies, queries the supplier for missing information, routes approvals, and posts the payment. The human reviews exceptions - let start conservatively with 20% of the volume. The rest flows through without anyone touching it.

Regulatory compliance triage. A compliance notice arrives from the regulator, an agent reads it, classifies the obligation, checks it against current controls, drafts the initial response, and escalates with a recommendation. The legal team reviews and signs, the grunt work (reading, classifying, mapping) is done.

Cloud cost optimisation. The AWS bill spikes, an agent detects the anomaly, traces it to a misconfigured instance, evaluates whether reserved pricing or a different instance type fixes it, applies the change, and logs the decision. Finance gets a weekly digest, nobody gets paged at 2 AM. Granted this is the more complex, but give your AI Engineer grad the right business support and she can knock this out for you - just don’t start here.

These aren’t science projects. Every piece above exists in production today at companies you’ve never heard of, the difference is they thought in job-shaped units, not feature-shaped ones.

I’ve been reverse engineering the deployments that actually deliver value, and there’s a pattern, put simply boundary clarity is everything. The agent owns a defined scope with explicit escalation rules. “This agent handles invoice reconciliation for suppliers in APAC. If the discrepancy is over $10,000, it stops and asks.” Not “this agent is smart and will figure things out.”

The human sits at the decision point, not in the workflow. You don’t need a human to approve every step, you need them at the inflection points: the exceptions, the judgment calls, the things that would cost more to automate than they’d save, the agent does the other 80%.

Share

And you measure closed-loop throughput, not usage. Nobody cares how many prompts the agent processed. The metric is: “of 1,000 invoices that arrived this week, how many reconciled and paid without human intervention?” Throughput → Completion → Error rate, the same metrics you’d use for a person.

I’ve also seen the ways this goes wrong. The one agent to rule them all trap: someone builds an “enterprise AI agent” that does everything, and it fails at scope. The domain is too broad, the edge cases are infinite, and nobody trusts it. The winning approach is narrow and deep, one job, done well, it’s not sexy but the results are credibility building and that is!

Then there’s the platform before jobs trap, someone buys an agent platform before they know what jobs they’re automating, they end up with a very expensive dashboard and nothing that actually runs. Pick the job first, then pick the tool.

And the autonomous equals unattended trap, an agent that runs without oversight is called a liability. An agent that sends a weekly summary of what it did is called an employee. Design for auditability, not autonomy, trust is earned in the log file.

Next time someone pitches you an AI strategy, ask them one question: what’s the job you’re going to automate? Not the task, the job.

If they describe a chat window, they’re selling you a copilot. If they describe a dream, they’re selling you a keynote. If they describe an end to end process with clear boundaries, escalation rules, and measurable throughput, you’ve found the agentic middle. That’s where the value lives. Between the feature and the fantasy. And it’s available right now to any organisation willing to think about work differently.

The other day I argued that the sovereign model conversation is a trap. The risk it responds to; geopolitical friction narrowing access to frontier AI is real. But the solution is the most expensive, slowest, and least effective hedge available for ANZ firms, the same applies at a country level. So here’s what actually works.

The alternatives I’m going to walk through share a common theme, none of them require a data centre, none require training a foundation model, mone require a team of ML researchers. They are procurement decisions, architecture patterns, and operational disciplines that any ANZ business can implement on a realistic timeline and budget. Not all of them are right for every firm, but one of them is right for yours.

Option 1: Multi-model routing

The single most effective thing you can do to reduce AI supply chain risk costs almost nothing in new infrastructure.

Stop treating an AI provider like a platform and start treating it like a commodity supplier. Design your integration layer so that switching models is a configuration change, not a rebuild. This means:

• Wrap model calls behind an abstraction layer (OpenAI-compatible API, LangChain-style orchestration, or your own thin adapter)

• Never hardcode model names, endpoint URLs, or API keys in application code

• Run the same workflow against at least two providers in parallel during testing

• Include a fallback provider in your production deployment

Why this works: Model agnostic architecture is good engineering practice regardless of the risk question and it gives you optionality at almost zero marginal cost. It also forces your team to treat model quality as a measurable property (which it should be) rather than a platform bet.

Who it’s for: Every firm that uses a third party model API. There is no excuse not to do this.

Simplest way to test: Try openrouter.ai (not an endorsement)

Option 2: Open Source ops readiness (the insurance you keep paying for)

I’m not suggesting you run Llama or Mistral in production for every workflow tomorrow. But I am suggesting you build the operational muscle to do so for a subset of your workloads, tested and documented, within your current budget cycle.

This means:

• Pick one non-critical workflow that could tolerate lower model quality

• Deploy an open-weight model (Llama 4, Mistral Small, Qwen, DeepSeek) via your cloud provider’s managed inference (Bedrock, Vertex AI, SageMaker)

• Measure the gap in quality, latency, and cost versus your primary provider

• Document the runbook: deployment, monitoring, scaling, rollback

Why this works: The operational capability—deploying, monitoring, and managing your own model instance—is worth more than the model itself. If you have the ops muscle now, you can scale it when geopolitics forces your hand. If you don’t, you start from zero at the worst possible time.

Who it’s for: Firms with existing cloud infrastructure and a DevOps function. Budget: a few weeks of engineering time plus inference compute costs.

Simplest way to test: get your dev’s to spin up some POC’s in Azure Foundry (not an endorsement) you’ll be up and going in an afternoon.

Option 3: Cloud abstraction (the easy default)

If multi-model routing is the cheapest hedge and open-source ops is the proactive one, cloud abstraction is the one you probably already have without realising it.

If you’re buying AI through AWS Bedrock, GCP Vertex, or Azure Foundry, your cloud provider has a commercial incentive to absorb regulatory friction on your behalf. When Anthropic or OpenAI changes terms, the cloud provider negotiates. When export controls shift, the cloud provider routes around them, or finds a compliance path.

Why this works: Cloud providers are repeat players with legal teams, compliance infrastructure, and commercial leverage. They also have multiple model vendors in their marketplace, so switching between providers is often a few clicks. You lose some control but gain operational simplicity and institutional mediation.

Who it’s for: Firms that already have a cloud relationship and want to solve this without adding architectural complexity. The cost is whatever markup your cloud provider charges on model inference, generally 0-20% over direct API pricing.

Simplest way to test: Your AI Engineering team will already be doing POC’s in this space, have them explain in business terms what they’re doing.

Option 4: Do nothing responsibly (the valid non-answer)

The final option is the one nobody will say out loud: accept that the current regime is stable enough for your planning horizon. This is not denial, it’s an explicit risk acceptance decision, documented in your risk register, with a review cadence.

What this actually requires:

• A one-page risk note in your existing IT risk framework: what specific event would trigger a pivot, what the impact would be, what the warning signs look like

• A [bi]annual review tied to your strategy cycle

• An agreement that if warning signs hit a defined threshold, you escalate to one of the three options above

Why this works: Many ANZ firms genuinely don’t face enough exposure to justify even the modest costs of options 1-3. If your AI usage is through SaaS products (where the vendor manages the model relationship), your supply chain risk is indirect and deferred. Building model-agnostic infrastructure for one summarisation workflow is not a good use of capital.

Who it’s for: Firms with limited direct AI API exposure, heavy reliance on SaaS-based AI features, and shorter strategic planning horizons. The key is making the decision explicit rather than accidental.

Simplest way to test: do nothing but the above.

How to choose

The decision framework is simple.

The real insurance isn’t a model! Here’s the conclusion I’ve arrived at after arguing with myself about this for two weeks:

The firms that navigate AI supply chain risk best won’t be the ones that build their own models. They’ll be the ones that build the capability to switch, that capability (model-agnostic architecture, operational readiness, supplier diversity) is valuable regardless of whether the geopolitical risk materialises. It makes you a better buyer of AI, a more resilient operator, and a clearer-eyed strategist.

The sovereign model conversation was a useful trap to argue through, but the right answer for ANZ firms isn’t to build it, it’s to build optionality.

I’ve now had four conversations in as many weeks where an ANZ executive tells me they’re “looking into” a sovereign AI model. The trigger is always the same: a vague but genuine anxiety about geopolitical risk. What happens if the US restricts model access? What if sanctions shift? What if we build half our technology stack on someone else’s API and wake up one morning with no fallback?

These are good questions, but the answers most people reach for are based on fear, not on what the rules actually say.

There is a real, documented, evolving regulatory framework that governs this stuff. It’s published by the US Bureau of Industry and Security. It’s called the Export Administration Regulations. And almost nobody I talk to has read it.

The sovereign model pitch sounds compelling: train a model under local law, on local data, with local governance. No US export controls. No Chinese sovereignty questions. No dependency on foreign labs.

Here’s what it costs to make that real:

• Capital. An Apertus scale project runs tens of millions. Frontier grade training runs hundreds of millions. Even a fine-tuned open-source deployment with serious compliance overhead costs seven figures before it delivers value.

• Talent. The people who can train frontier models are employed by the frontier labs. Australia has maybe a few hundred AI researchers with relevant experience, New Zealand has fewer. Building a sovereign capability means competing with Google and OpenAI for that pool. If someone thinks you can do it without this capability i’d politely thank them, put down the coffee and end the meeting.

• Time. Apertus took years, by the time your sovereign model is ready, the frontier will have moved twice, you’ll be building yesterday’s insurance for tomorrow’s risk.

• Ongoing cost. Model maintenance, data pipelines, evaluation, safety testing, deployment infrastructure, a sovereign model is not a build once asset, it’s a permanent operational commitment.

And after all that: your sovereign model still won’t be as capable as the frontier models your competitors have access to. That’s not a failure of execution. It’s a structural reality of the economics of AI training. This is like building your own power plant because you’re worried about grid reliability. Technically a hedge. Practically insane unless you are the grid.

The real question ANZ firms should be asking

The concern driving all of this, supply chain risk from geopolitical friction is valid, the response (sovereign model) is not.

The gap between the risk and the response is where the real value is. There are cheaper, faster, more adaptable ways to reduce AI supply chain risk that don’t require a data centre or a foundation model of your own. They just require a different mental model of what “insurance” looks like in AI procurement.

This piece has laid out the problem and the false solution, in Part 2, I’ll walk through the four alternatives that actually work for ANZ firms, and how to choose between them.

I think the arrival of AI coworkers and stronger copilots changes the economics of work more than most organisations are ready to admit. For a while, the conversation sat around software subscriptions, productivity uplifts, and whether a given model was slightly better than the last one. That framing now feels too small. Once a digital coworker can take on meaningful slices of research, drafting, synthesis, coordination, and review, the real question is not what the licence costs. The real question is what happens to the cost structure of the business built around that labour.

That is where things start to get interesting. Historically, knowledge work has had a stubborn shape to it. If you wanted more output, you usually needed more people, more specialist time, more management overhead, and more delay built into handoffs between teams. Software could improve the process, but the human cost base still dominated. AI coworkers change that equation because they introduce a new kind of variable labour. It is not labour in the legal or human sense, but economically it behaves like a new pool of work capacity that can be turned on, directed, reviewed, and scaled at a cost that looks very different from traditional headcount.

That does not mean staff disappear or that businesses become empty shells run by prompts. I do not think that is the serious story, the serious story is that the cost curve for certain kinds of work starts to bend. Tasks that were once too expensive to do thoroughly, too slow to do consistently, or too operationally annoying to prioritise can suddenly become viable. Internal briefings can be tighter, research can be broader, drafting can happen earlier, scenario analysis can happen whilst developing the idea. Documentation can happen! The business can afford to think more before it acts, or at least it can afford to test more options before committing to one.

That sounds positive, and in many cases it will be. But it also creates management pressure. If the cost of producing competent first-pass work drops sharply, then the bottleneck shifts somewhere else. Usually it shifts to judgement, workflow design, governance, and prioritisation. In other words, the more capable the copilots and coworkers become, the less the business problem is about generating content or analysis and the more it becomes about deciding what work should exist, what quality standard matters, who signs off, and where human attention is genuinely valuable.

That matters for the economics of business because many firms still carry cost structures built for older assumptions. They assume valuable work is scarce because human production capacity is scarce. They assume managers must control flow through slow approvals because generating new work is expensive. They assume expertise must be protected in narrow channels because accessing it is limited by time. AI coworkers start to challenge all of that. When the cost of producing a decent draft, a decent synthesis, a decent options paper, or a decent first-pass recommendation drops, the economic logic of the organisation starts to change with it.

NZD15 to save a day’s effort developing a project plan from 10 different data sources and business context: This is token economics!

I think that creates at least three major implications for business. The first is margin pressure in any industry where value has been partly sustained by expensive knowledge-work throughput. If more firms can produce reasonable strategic, creative, analytical, or operational work at lower cost, then old pricing assumptions become harder to defend. The second is an internal productivity reshuffle. Teams that learn how to combine human judgement with AI-generated first-pass work will likely move faster and cheaper than teams still organised around manual effort at every stage. The third is a management reset. Leaders will have to become much clearer about what work deserves expert human time and what work should be delegated to a machine-assisted layer by default.

That is why I think the phrase copilot is becoming a bit too modest. A copilot sounds like a supportive utility sitting at the edge of the workflow. A coworker sounds closer to what many firms are actually buying. They are not just buying autocomplete. They are buying parallel work capacity. They are buying more shots on goal (tried and failed not to bring in a football analogy during the world cup), they are buying faster internal loops, they are buying the ability to run more tasks without adding equivalent payroll cost. Once you see it that way, the economic implications become much harder to ignore.

There is a catch though, lower unit cost for knowledge work does not automatically create a better business. In some organisations it will just create more noise, more reports, more drafts, more experiments, and more management confusion. If the operating model is weak, AI can flood the system with activity that looks productive without producing real advantage. That means businesses may need to redesign workflows at the same time they adopt these tools. Otherwise they risk creating an economy of abundant output and scarce judgement.

I also think this will create a split between firms that treat AI as software spend and firms that treat it as operating leverage. The first group will debate licences, pilots, and budget categories. The second group will ask harder questions that matter:

• Which work can now be done at far lower cost?

• Which internal services should be rebuilt around that fact?

• Which roles become more valuable because they provide judgement, trust, and decision clarity?

• Which parts of the business should now be smaller, faster, or more experimental because the economics have changed underneath them?

The most interesting implication may be that business strategy itself starts to change. When the cost of thinking, drafting, analysing, and coordinating comes down, the viable shape of a firm changes too. Smaller firms can act bigger. Larger firms can become less excuse-driven about slow execution. Niche operators can punch above their weight. Middle management may need to justify itself less by control and more by decision quality. Specialist knowledge may become more amplifiable. In that environment, the winners will not simply be the firms with the best model access. They will be the firms that understand the new cost economics quickly enough to redesign how work moves.

That is the part I think many leaders still underestimate. AI coworkers are not just another tech layer. They are a change in the economics of capability. And once the economics change, the business eventually has to change with them.

A clean overhead shot of a single A4 page on a wooden desk beside a closed MacBook. The page is handwritten with five short notes under headings like “what I asked, what it gave me, what I decided”. Soft daylight, black and white. The artefact at the centre of the article, photographed honestly.
Wendy Folau-Kennedy published something last week that I have not been able to stop thinking about. Her article “Provenance Is Not the Point” reads the Deloitte hallucination fiasco and the KPMG resignations as the same failure in different costumes. The records were fine. The thinking was missing. She calls the new bad habit the Provenance Reflex: knowing where information came from gets confused with engaging with what it says. Most coverage is treating this as a Big Four problem. I think she has actually written the warning label for every SME owner in New Zealand and Australia.

The numbers explain why. Datacom’s 2024 survey put NZ business AI use at 67%, up from 48% the year before. MYOB analysis released this year says Australian SMEs using AI are growing 2.8 times faster than those that aren’t. Around 60% of Australian small firms are using or planning to use AI. Adoption is racing ahead of any habit of recording what the AI did, what the owner thought about it, and what they decided.

Here is where Wendy’s framing translates badly for SMEs, and where I want to challenge my own earlier writing on everydayAI. I have argued before that a clear AI policy is the right first move for a small business. After reading her piece, I am revising that, a policy is a document, the Reflex is a behaviour, documents do not interrupt behaviours. The Deloitte team had every policy template in the country. They still shipped fabricated citations.

The four bad habits Wendy names in the original piece (summary substitution, consensus capture, disclosure as defence, and policy theatre) show up in SME form every day. Summary substitution is the owner who feeds a 40-page commercial lease into ChatGPT and signs based on the bullet recap. Consensus capture is the founder who asks Copilot whether the new pricing is competitive, receives the agreeable answer all language models give, and walks into Monday’s team meeting convinced. Disclosure as defence is the line at the bottom of a client proposal saying “AI was used in preparing this document”, as though the sentence transfers responsibility to the reader. Policy theatre is the AI policy downloaded from a template site, signed off by the board of two, and never opened again.

The legal exposure is where SME owners need to pay attention. Section 137 of the Companies Act 1993 sets a duty of care for every NZ company director, including the one who is also the receptionist. In Australia the equivalent is section 180 of the Corporations Act 2001. Wendy points out, correctly, that the section 138 reliance defence largely collapses when you are the person who curated the AI input in the first place. Owner-operators are almost always in that position. The defence “the AI suggested it” does not exist. It has never existed.

There is a forward-thinking opportunity hiding in this that the big firms cannot grab. Small operators can produce the artefact Wendy says regulators will eventually ask for, faster and cheaper than any enterprise governance committee. One page per material decision. What I prompted, what the model gave back, what I considered, what I rejected, what I decided and why. Three minutes of writing. It is the closest thing to a section 138 defence a sole director currently has, and almost nobody is doing it.

The counter-opportunity is starting to show up in client work already. SME owners who include a short, plain “how we used AI on this” note in proposals are winning work from buyers who got burned by the Deloitte-style failure mode last year. Provenance plus engagement is now a sales asset rather than a compliance line.

Here is my possibly controversial alternative, and it cuts against most of the advice circulating. Do not try to keep an engagement record for every AI-touched decision in your business. You will give up by week three. Pick the five decisions a year where personal liability is real: hiring and firing, any contract above your pain threshold, financial forecasts you put in front of a bank, anything in a regulated category, and major pricing changes. Document those properly. Ignore the rest. Trying to govern every Copilot prompt is the new compliance theatre, and it will burn small operators out faster than it protects them.

Wendy ends her piece with a question I want to point straight at every SME owner reading this. If a regulator, a bank, or your largest client asked you tomorrow to produce evidence that you personally engaged with the AI inputs to your last three material decisions, what would you submit? Most would submit nothing, because there is nothing to submit. That is the gap worth closing this year. Full credit for the framing here belongs to Wendy Folau-Kennedy. Read her original piece “Provenance Is Not the Point” for the source thinking.

Dan Shipper of Every published an 8,000-word essay called After Automation on 21 May. I read it twice. I think it is the most useful single piece written on AI and work this year, and I agree with the spine of his argument. That is rare for me, so it is worth saying why.

The central observation is this. Every is about as AI-native as any company you can name. Codex, Claude Code, internal agents, AI responding to ninety-five per cent of his email. The team has grown from four people to nearly thirty since GPT-3. More automation, more humans. Shipper’s explanation is the line I have been circling for two years without quite catching it. Yesterday’s expert competence is now cheap. AI floods the zone with work that looks expensive and is mostly fine, which collapses the average value and pushes a premium onto judgment, taste and finishing. Experts become more valuable. They become system designers.

I have written here about the squeeze on middle management and the way agents expose process debt. Shipper’s framing makes those pieces sharper. The roles being hollowed are coordination and aggregation. The roles being amplified are the ones where someone has to decide what good looks like in a situation that did not exist last quarter. I am updating my own thinking. I used to describe this as a quiet hollowing. It is closer to a re-pricing. The same work that was expensive on Friday is cheap on Monday, and the premium has moved next door.

Now contrast that with ClickUp. A few hours before Shipper’s piece dropped, Zeb Evans cut twenty-two per cent of his workforce, replaced them with three thousand internal agents, and posted million-dollar salary bands as a flex. Gartner’s own data says eighty per cent of companies pulling this move do not see guaranteed financial returns. ClickUp is not a bold AI restructure. It is a SaaS company having a difficult year and dressing the cuts in something more flattering. When a CEO announces record results and a twenty-two per cent layoff in the same breath, that is a tell. Nvidia’s Jensen Huang has called the narrative that links AI to job loss “too lazy”. He is correct.

The credibility problem at the AI lab CEO level is worse. Dario Amodei warned of a white-collar bloodbath, predicting half of entry-level roles could vanish in one to five years. He has since reframed automation as a multiplier of output. Sam Altman told the Commonwealth Bank of Australia he was “pretty wrong” about the economic impact. These reversals are happening as both labs prepare for trillion-dollar IPOs. Treat their forecasts as marketing.

The more honest voice is Ethan Mollick at Wharton. He told the CNBC Workforce Executive Council Summit that “no one knows anything” about AI’s impact on jobs, including the top AI research labs. Nobody has five years of generative AI experience. Anyone selling a confident five-year forecast is either confused or selling something. Simon Willison’s posture is similar. He treats LLMs as tools to ride and report on, not prophesy about, which is the same instinct as Shipper’s “ride the models” advice in different clothes.

Then there is Ken Griffin of Citadel. Until last year he was a vocal AI sceptic. He now says AI is “real” and that he went home one Friday “fairly depressed” after watching agents do work he used to pay master’s and PhD analysts to do over weeks. That moment is the most dangerous one in any leader’s AI journey. The first time you see capable agents in your own domain, your instinct is to over-react and cut. Six months later, half of these companies are quietly trying to rehire.

Where I would push Shipper. “Ride the models” as universal career advice is necessary but not sufficient. Riding without taste turns you into a slightly faster source of the same generic output everyone else is producing. The premium is on the framers, the people who decide what to automate and where to over-invest human judgment.

The forward call is simple. Hire framers as well as doers. Pay for taste. Stop measuring productivity in tickets closed and start measuring it in decisions made well. The quiet opportunity is to become the person who builds the rails, review systems, evaluation suites and editorial standards. Almost no job description currently reflects that, which is exactly why it pays.

The controversial alternative. A meaningful share of AI-branded layoffs are not about AI. They are cost cuts at companies that were already bloated, badly managed, or had missed an earlier strategic shift. AI is the most flattering excuse on the market right now. If we accepted that, the policy conversation and the headlines would look very different. A few CEOs would have nowhere to hide.

introduction

The label is unfortunate. Hyper velocity engineering sounds like a 2003 management book, and the first thing my team would do is roll their eyes. But underneath the name is a free Microsoft repository, MIT licensed, that any New Zealand engineering team can fork tomorrow and start adopting without procurement gymnastics or consultant invoices. I think it’s the most coherent attempt I have seen at turning GitHub Copilot from an autocomplete into an actual engineering system.

Here is what I am seeing across mid to large NZ teams right now. Most have already bought Copilot or Cursor. Individual developers are coding faster. PR queues, test debt and security review are not, and the gap between the two is widening. Your investment is producing motion without flow, and the cost of that gap will eventually surface as a quarter where shipping velocity quietly stalls and nobody can explain why.

This two-part article is my answer. The free section you are reading sets up the problem and the opportunity. The paid section is a 100-day execution plan I would use myself, sized for both a 10 to 20 person product team and a 50 to 100 person engineering org. It tells you exactly which of the 221 HVE artifacts to switch on, in which order, against which lifecycle stages, and how to baseline the four metrics that prove the change is working. It also includes a 12-month arc, two counter-opportunities and one alternative that is genuinely uncomfortable to recommend.

The hardest part of this transition is not the technology. It is getting your team and your manager to agree to the first move without it being pattern-matched as another framework. I will explain the framing that has worked for me and the framing that quietly kills these initiatives in steering committees. If you walk away with one defensible plan you can put in front of an EM or CTO on Monday, with zero new vendor spend, I think the subscription is the cheapest piece of advisory work you will buy this year.

Consider a paid subscription, don’t pay a consultant for what will cost you a coffee a month

Strategy and execution plan (paid subscribers only)

Read more

I listened to Art Shectman, CEO of Elephant Ventures, on the Emerj AI in Business podcast this week, and his framing is the cleanest articulation I have heard of what is going wrong inside banks, insurers and asset managers right now. The headline takeaway is uncomfortable for anyone running a multi-year transformation programme. If you cannot put a workflow into production in 90 to 100 days, you are working on the wrong thing.

That sentence is doing a lot of work. The argument is about scope, not corner-cutting. The thinking I have written about previously on the activity-impact gap (around 80 per cent of organisations active in generative AI, around 60 per cent still reporting no measurable EBIT contribution) sits behind this. The reason so many programmes are stuck comes down to scope rather than model capability. A twelve-month programme will be obsolete on a tooling basis before it reaches production, and worse, the board will have lost patience. Shectman is blunt that boards are done with strategy decks. They want something alive in the business. It does not have to move the P&L yet. It has to exist.

Where my thinking has shifted since I wrote about the four-phase gate model earlier this year. I previously assumed that small wins were the on-ramp to bigger bets. I am less convinced now. The technology is changing every quarter. A leader who uses three quick wins to justify a twelve-month programme is trading optionality for ambition at the moment when optionality is the more valuable asset. The better posture is to keep shipping small things in chains. Each increment earns the right to the next. None of them is so large that you cannot pivot when the next capability shift arrives.

The second piece of the puzzle is vendor selection, and this is where most leaders are bleeding time. Shectman calls it pitch and platform exhaustion. Twelve pitches a week, three of them functionally identical, and a calendar that fills up faster than any meaningful evaluation can happen. Be deliberate about why you are in the room. Go to the hyperscalers and the household names when you want generalist education on what each layer of the stack does. They will give you a clean walk-through and a reusable vocabulary. Go to the startups when you have a specific niche problem inside your domain that the big players are not solving well. Their founders know things the mainstream pitch has not caught up with yet.

The discipline I want to add here is one Shectman touches on but does not name directly. When you find a small vendor with a sharp product, resist the urge to drag them into customising for you. The temptation is strong because they want your business and they will say yes. The better question is whether you should re-engineer your internal process to fit their product. That sounds counter-intuitive inside a regulated enterprise, but it preserves the vendor’s roadmap and avoids the slow-motion failure that follows when you bend a small company into a shape it cannot maintain. If your process genuinely cannot bend, you have learned something useful. This is not the right vendor and you should walk.

The third thread ties the others together. Action compounds. Shipped work earns permission. A leader who can point to one production deployment in front of the board on Monday morning, even an unglamorous internal one, is having a fundamentally different conversation than the leader presenting another roadmap. Shectman calls it the grease the wheel needs. The institutions making genuine progress are the ones that have stopped trying to be ready and started shipping.

A possibly controversial position. The institutions that will outpace their peers over the next eighteen months are the ones that publicly close their door to new vendor pitches for a fixed window, perhaps six months, and commit to a three-vendor harness. Constraint produces shipped work. Open evaluation produces meetings. There is a respectable argument that a public commitment to fewer vendors and more delivery is itself a strategic signal, both to the board and to the market. It is the opposite of what most procurement teams have been trained to do, and that is precisely why it might work.

The counter-opportunity is worth naming. Some regulated workflows should not be sprinted at all. A small number deserve deliberate friction, including shadow-mode operation for six months or longer and named human review gates before anything goes live. The skill is knowing which workflows belong in the 100-day bucket and which belong in the slow lane. Most leaders, in my reading, are still putting too many in the slow lane.

The numbers in the strategy work I’ve been reviewing tell a quiet but uncomfortable story. Around 80% of organisations now use generative AI in at least one function. About 62% are already experimenting with agentic AI. And roughly 60% have still seen no enterprise-wide EBIT impact from any of it. That third number is the one that should keep boards awake. Activity has been mistaken for impact, and the gap between the two is getting wider.

Most AI inside the average company today is horizontal. Chatbots, copilots, summarisers, meeting note-takers. These tools help individuals work faster, and that feels like progress. The problem is that faster individual work rarely changes a P&L. A salesperson who drafts emails in half the time still closes the same deals. A support agent who summarises tickets in seconds still works the same queue. The work is lighter. The number at the bottom of the page is the same.

What separates the companies starting to see real movement from the rest is not a better model or a bigger budget. It is three habits applied with discipline. Value is defined up front, with a metric at every layer from technical performance through to financial impact, before a single line of code is written. Measurement is wired into the rollout itself, using A/B tests or staggered deployment so attribution is clean from day one. AI is run as a managed investment. There is a fixed review cadence, a clear stage gate at the end of each phase, and one evidence pack per use case that tracks benefits and total cost of ownership in the same document. Only the use cases that prove value advance.

The framework I’ve been working with sits in five layers and is read bottom-up. Technical performance, owned by data science and engineering, covers hallucination, latency, token cost and drift. User adoption sits above it, owned by product and frontline ops, and measures daily active users, workflow penetration and how often people accept the AI’s output versus override it. Operational KPIs come next, owned by a named end-to-end process owner, and track cycle time, defect rate and cost per case. Strategic outcomes sit above that, owned by the business unit GM. Financial impact, owned by Finance and FP&A, is the layer everything else has to credibly support. No layer is allowed to advance a use case if it cannot hold up the one above it.

This is where the forward-thinking opportunity sits, and it is less obvious than it sounds. The fastest way to kill bad AI use cases early is to give Finance the keys, not the data team. FP&A is paid to be sceptical about unit economics. If they own the financial layer of the framework from week one, the conversation about cost-to-serve, margin and total cost of ownership starts before the proof of concept burns six months and a million dollars. The data team builds. Finance gates.

The counter-opportunity, and the one I’d push hardest, is to stop chasing enterprise-wide rollouts altogether. Pick three to five end-to-end workflows where a single owner can be held accountable for a single financial outcome. Order-to-cash. Claims handling. Onboarding. Field service dispatch. These are the places where AI changes the shape of the work itself. Everything else can wait.

Now the controversial bit. I think most internal copilots should be sunset. They are a productivity tax dressed up as an asset, and the licence spend is rarely justified by anything measurable. The money would do more work redirected into a smaller portfolio of agentic, workflow-changing use cases that can be measured like a capex project. That is going to be an unpopular view in companies that have just signed three-year enterprise agreements, and I understand why. But the evidence pack on most copilot deployments is thin, and the honest answer to the gate question of “is there measurable operational and financial impact that justifies scaling” is usually no.

This builds on the piece I wrote earlier about the C-suite and the frontline closing in on the messy middle. That article argued the squeeze was already on. The piece I left out, and which I’d update now, is the scaffolding. Without a five-layer measurement framework and a hard stage gate at the end of every phase, the squeeze produces noise instead of signal. Leaders get the activity they asked for. They do not get the impact.

The next step for any team reading this is simple. Agree the five-layer metrics and the stage-gate cadence for your top three AI use cases this quarter. Then run the cadence. Monthly review. Quarterly gate. One evidence pack per use case. The companies that do this are going to look very different by the end of next year from the ones still talking about pilots.