In October 2019, I was asked to host the SurveyMonkey marketing leadership offsite in my backyard. I said yes on one condition: we were going to have a lip sync celebration. I rewrote the lyrics to Macklemore’s “Thrift Shop,” recruited the leadership team to perform first in full costume, and applied enough cheerful, relentless peer pressure that people showed up with choreography and props. People still call that night one of the highlights of their careers because of the pure joy. What they felt in the room was designed in the quiet moments before anyone walked through the door.

Design for the feeling

Maya Angelou observed that people don’t remember what you did or said, but how you made them feel. At the Bay Area Lovable SheBuilds event, tactics were emailed in advance so that when I invited volunteers to arrive early, we could focus entirely on the feeling we were there to create: belonging and confidence. When I closed the celebration by asking a room of strangers to sing and dance to Sia’s “We Are Unstoppable,” a dozen women joined me on stage without hesitation. What started as polite, uncomfortable smiles became singing and dancing because a small, prepared group made it safe for strangers to follow.

Give before you get

I am hosting an executive dinner with a client this week, and the goal, as it is for most events like this, is to create relationships that could lead to new business. It’s easy to slip into designing an event around what the host wants, but meaningful relationships come from starting with what you are giving: thought-provoking conversations, camaraderie, and connection. Before I design any gathering, I think about what people are trading off to be in the room. I read LinkedIn profiles from the bottom up to understand the journey that brought each person to today because connections within a group always begin with connections one on one.

Set the rules, then hold them

Rules of engagement for an event might sound confining and yet by setting cultural norms, you create the freedom for people to take risks they wouldn’t otherwise take. A neighborhood gathering I attended for the purpose of celebrating learning and mediocrity requires every attendee to perform so no one has to be brave alone. At every SandboxTime AI session I run, my opening rule is drop the shoulds and embrace that during our time together we will advance our skills. Unlike sessions where you watch others build, everyone builds and everyone shares. The rules make it safe to be uncomfortable together and that is where the learning actually happens.

Whoever holds the mic goes first

When I hosted the SheBuilds event, I opened by confessing to 250 women that I had spent months as an AI lurker: watching, reading and encouraging others to build while embarrassingly not building much myself. What finally tipped me was Oli, an 88-year-old woodcarver I met in line at the post office who didn’t have a website for his handmade cutting boards and so I built him one. I told the room that if I could host this event two months later, they could build something in an hour. When the person holding the mic says “I am still figuring this out,” everyone else gets permission to stop performing and start learning.

The quiet before the magic

The most powerful antidote to the anxiety AI is generating is being in a room with people focused on learning together. What makes that room possible is the balance of doing and thinking in the preparation with logistics handled in advance so there is space to focus on the feeling of the experience.

This year I have had the privilege of bringing these principles into a women’s AI workshop for over a hundred women and a company offsite built around accelerating AI capability. If you are looking to bring together prospects, customers, or teams and want to think through what that could look like, I would love the conversation. You can learn more about how I work at christinerimer.com.

While magic happens in the room, it is created in the quiet moments before.

The term anniversary implies celebration and reflection, be it a marriage, a life, or a career. I’ve never considered a subscription renewal an anniversary until Claude Pro. My first thought was embarrassment that I only met Claude a year ago, as it seems so much longer. I suspect I’m not alone in feeling like I should have started sooner and moved faster. That said, the milestone has caused me to reflect on my journey.

Phase 1: Getting to Know You

When I transitioned from ChatGPT to Claude Pro, I was still engaging like a search engine with isolated conversations: ask a question, get an answer, and start over next session. Claude didn’t know me and I didn’t know Claude. Whenever I hear AI cynics say that AI’s answers are generic, I expect they are in these early days and have yet to invest in providing Claude the context to leverage the capabilities.

In the early days, I downloaded the desktop version which is organized into three tabs: Chat, Co-work, and Code. My first shift was to start to explore Projects which are accessible from Claude’s desktop, mobile and browser version. Projects are folders for a recurring set of chats or conversations where you can add logic or instructions guiding that set of work. Today I have projects for my work such as AI Angles articles, Sandbox Time AI events, and each consulting client. I also have personal projects such as my backyard redesign and my closet refresh with photos of my favorite clothing pieces, style preferences, and chats to help me fill in the gaps for specific occasions since I hate to shop. The shift from chat to project sounds small, but it’s a fundamental mindset shift to build context with each conversation and create repeatable capability beyond today’s task.

In Co-work, I connected the applications I use every day such as Google calendars, Drive, email, Chrome, Granola, Gamma, and Slack to lay the groundwork for addressing daily workflows with agents. I started conservatively with approvals each time before getting more comfortable with what it meant to have a tool inside my actual working life, helping me manage my time and attention.

Phase 2: Building Together

My relationship with Claude evolved when I created a voice profile inspired by Ruben Hassid. Ruben’s prompt is a multi-hour interview about who I am, how I write, what I hate, and what I believe. The output is a document called Voice_Profile_Christine.md, which I add to any Claude project as a reference. Claude knows I hate em dashes, short dramatic sentences, and filler setups. Claude knows I anchor advice in personal experience, prioritize being practical over preachy, and want to inspire people to take action. The insights Claude played back after each interview question shifted my relationship from asking questions to inviting thought partnership.

In my AI Angles project, I partnered with Claude to write the instructions for writing articles. I share my idea by dictating a draft that includes the opening hook concept, my personal experience, and top lessons or takeaways. Before writing, Claude reads my latest voice profile and asks clarifying questions. I run tag-team rounds of edits, and when each article is complete, I ask Claude to review all the edits I made, identify new patterns or rules, and fold them back into the voice profile for next time. The partnership allows me to lean into my strength as a storyteller to create connection, build community, and be a catalyst for readers to take action.

This phase is also when I launched Sandbox Time AI. Because I work better together than alone, I set it up as a forcing function for building with AI the same way a gym’s group exercise class does for working out. After experimenting with sessions on Lovable, Gamma, and ChatPRD, I’ve focused these sessions on Claude to build my own skills while inspiring others to do the same, an hour at a time.

Phase 3: Creating Katniss

When I decided to build my first agent, I was inspired by Claire Vo’s How I AI episode with Jesse Genet to give her my agent a name. I chose Katniss, the Hunger Games character who represents the scrappy, resourceful, and collaborative relationship I wanted to create with my virtual Chief of Staff. Inspired by my friend Annie Tsai, I wanted Katniss to reduce the cognitive load of managing multiple calendars across personal, professional, and client commitments. While I got Katniss up and running in Claude Code, I haven’t yet integrated her into my daily routine and temporarily “killed her” by reorganizing my desktop folders. I’m working through whether Katniss should live in Code, Co-work, or as a live artifact, and how best to work her into my daily workflow of checking calendar, email, Slack, and my Google Doc launchpad. Inspired by Claire Vo’s How I AI session with Felix Rieseberg, I’ve just started exploring live artifacts by building a weekly dashboard that can be easily redesigned and refreshed.

You Are Not Behind

Wherever you are right now is the starting point for what’s next. Build out a project with instructions and sample files. Create a voice profile. Connect your applications. Check out live artifacts.

In my first group guitar lesson last year, our teacher Brandon Robinson shared that the goal was to play a new song together every week. We started with the two-chord “Horse With No Name,” and since then my binder has significantly expanded to include this week’s more challenging “I Get By With a Little Help From My Friends.”

The key to progress building with AI is the same as guitar: practice 20 minutes a day, embrace incremental learnings that will lead to transformative outcomes and build in forcing functions and accountability with others. Year two starts now. What’s your one move this week?

References:

• Women Defining AI for amazing courses and content

• Claire Vo How I AI with Jesse Genet on Building Agents

• Claire Vo How I AI episode with Felix Rieseberg on Live Artifacts

• Annie Tsai on The business side of AI

• Ruben Hassid on building a voice profile

• SandboxTime AI on My Claude Events

• My guitar lessons with Brandon

UCLA Graduation Day - a long time ago

Dear College students & recent graduates,

I have been thinking about you a lot lately. Beyond the AI news cycle, maybe it’s because my kids will graduate from college in 2027 and 2028. It could be because many of my friends’ smart, talented kids who graduated last year are still searching for full-time work. Finally, it might be because I started my own career in a recession when nobody was hiring and I have not forgotten the frustration. I’ve been thinking about what action you can take to navigate this moment and I keep coming back to these four skills you can hone that I believe can help.

1. Own the Outcome

Owning the outcome is a leadership principle I picked up during my days at Intuit. It starts with being really clear of your goal you are trying to achieve, designing a plan and then following through. While it sounds simple, being thoughtful about what you are trying to achieve and following matters even more when leveraging AI. Being clear about your point of view takes focus and attention. Asking AI to challenge you, find the holes, refine your thinking, read the responses, and refine again and again takes time. This article started with a thoughtful first draft and went through ten rounds of revisions with Claude before I deemed it “good enough” to publish. I set up Sandbox Time AI to own the outcome of my own AI learning by facilitating sessions and teaching others. For you, consider opportunities to “own the outcome” outside of a class or the scope of a summer job or internship. It could be helping your guitar teacher build sheet music with Claude or helping a small business owner neighbor build a website in Lovable. The irony is that it’s not about the actual outcome but rather practicing the end to end ownership.

2. Autonomy

The second skill is autonomy which means the capacity to govern yourself and make independent decisions. It is being proactive, self-motivated and self-disciplined. Companies are hiring people who proactively learned AI skills on their own outside the college curriculum. While I’ve paid for Stanford continuing education classes, bootcamps, Lenny’s Product Pass and Claude Pro, there’s no shortage or free AI classes, podcasts and YouTubes to learn and practice these skills. Like learning to surf or play guitar, you get good by putting in the reps while no one is looking.

3. Resilience

High-achieving students are well practiced in studying hard and getting the grade. However, the skill of bombing a test and digging yourself out, of struggling with a hobby, club or sport and pushing through rejections and challenge is the skill that build grit. This matters even more in an AI work world, because building with AI is fundamentally iterative. You learn by trying, failing, throwing away what you built, and trying again. The instincts are developed through time and repetition. My friend Chett Garcia, one of the top recruiters in the San Francisco Bay Area, has written about how founders still chase candidates from brand-name companies and miss the point. What he and I both keep coming back to is that grit and the willingness to keep going often beats the pedigree.

4. Storytelling

While it is widely known that many people fear public speaking, a 2023 poll found that about a third of Gen Z felt less confident speaking in public than they did before the pandemic. That makes sense, because surviving the terrifying middle school presentation is exactly the kind of formative experience that builds a foundation and this cohort lost some of those years. My own confidence in front of an audience started in sixth grade with school-wide poetry recitation contest. Being able to communicate your ideas, tell a story, and connect with an audience or your teammates is such an important part of being successful in the work world. This skill is not reserved for the select few who give big conference presentations or TED Talks, but rather for everyday moments: presenting a project, prepping for an interview, or speaking up in a team meeting. To practice, take opportunities to toast a group, teach something you love to a few friends or present an AI project you built. Being able to engage, excite and influence others your ideas and perspective through storytelling is a skill worth investing in.

So what

After six months of rejections when I graduated from college, I was desperate for a chance to prove myself. When a startup called Macromedia offered me a data entry clerk at $19.5K per year, I was profoundly grateful and on fire with the desire to demonstrate what I could do. Every chapter since has been built on those skills: end-to-end ownership, proactive autonomy, resilience and storytelling.

I invite your to practice these skills in a way that actually fits your personality and how you work. If you do well with accountability, find a partner, sign up for a class, join a cohort, and treat it like a group exercise class where the structure is what gets you in the door. If you do well alone, pick something you genuinely care about and build something in AI and share it presentation style with friends.

“Life moves pretty fast. If you don’t stop and look around once in a while, you could miss it.” Yes, I am quoting Ferris Bueller’s Day Off from 1986, which is older than you. I do not want to be flippant about how overwhelming this moment can be because the pace of change in jobs and economics and skills is real. I have a lot of empathy for what you are walking into. What my experience is taught me is to focus on what I can control and that’s my own time, attention and perspective. I believe practicing these skills can help open doors and I will be cheering for you the whole way.

Cheers,

Christine

References:

• Auren Hoffman article post on the job market and personal responsibility

• Chett Garcia, LinkedIn post on grit and hiring

• Fortune article on higher education rewriting the rules around AI

• Gen Z Public Speaking: Survation / Speak With Impact, 2023

This is my family in 2011 at our local Tahoe beach proudly sharing the sandcastle we built. The photo reminds me of those days of playing with the kids without distractions or goals. Spending time together on the beach playing was the point.

My kids are in college now. As we all grow up, we become more focused, busy, productive, outcome-oriented. It becomes harder to remember to spend time playing, experiment just for the sake of building and learning. I believe this is the mindset killing people’s ability to build with AI.

I launched SandboxTime AI in January 2026 as an experiment in giving that mindset back to adults, including myself. While there is no technical barrier to building with AI today, there are two less obvious but important barriers to overcome: blocking the time to play and experiment, and shifting the mindset from expecting to be an expert right away.

Whenever we learn something new, whether it’s guitar or surfing, we expect to put in the time. We know we need time in the water before catching a wave and hours practicing chords before playing songs around a campfire. Yet I watch so many people, myself included, struggle with that same mindset when it comes to AI. I want the productivity gains without doing the reps. The dialogue out there is so biased toward all the gains you can get from AI, which is real, but those gains don’t show up unless you put in the hours to develop the skills.

I named it Sandbox Time AI on purpose to invoke a childlike sense of play and curiosity, a medium where you build something with the mindset that it’s the time building that counts, the learnings you take away. What you build is likely to wash away when the hour is up.

Of course, knowing you should spend time playing with the tools and actually doing it are two different things. When it comes to my workout routine, I have weights, a rower, and an exercise bike at home, and I still pay for a CrossFit membership and commute to the gym four times a week. I go to Advanced because it’s dedicated space and time to work out, a coach pushes me through more reps than I’d ever do on my own, and watching my gym friends push themselves makes me push myself. Sandbox Time sessions run on the same logic, just with prompts instead of pull-ups.

Five Learnings from Sandbox Time AI

1. Lenny’s Product Pass vs. Claude: My original intention was to use the sessions as a forcing function for myself to try the products in Lenny’s portfolio. While I did run Lovable, Gamma, and ChatPRD sessions, I’ve increasingly leaned into Claude to develop my Cowork and Code skills. That’s where I’m spending my time practicing.

2. Free vs. Paid: All Sandbox Time AI sessions to date have been free, with the goal of helping, supporting, and collaborating with people as they build. The challenge with free events is the no-show rate, which is minimal impact for virtual sessions but a real one for in-person, such as the Lovable SheBuilds event. As my passion for being a catalyst grows, I’m being more thoughtful about how I balance free community events with paid offerings.

3. One-off vs. Cohort: Every session has been a one-off, a drop-in two-hour session with the same agenda: 30-minute intro, an hour of building, and 30 minutes of sharing. Some sessions sparked a real sense of camaraderie and a desire to stay connected, which is the gift of a cohort. For that, I’m leaning into Women Defining AI (WDAI) as my recurring learning program, which is a hundred times better than anything I could build solo. Go, Helen and the WDAI team!

   https://www.womendefiningai.org/

4. Virtual vs. In-Person: There are real challenges to in-person events, from finding a free venue with power and strong enough Wi-Fi to managing the no-show rate. I miss the in-person magic, though, so I’m testing two in-person sessions in the San Francisco Bay Area and Peninsula. https://sandboxtime.ai/events

5. Individual vs. Corporate Programs: Every session has been direct to consumers through LinkedIn. I’ve had early conversations with companies about running a Sandbox Time program inside their organizations. I consistently hear that organizations are introducing tools and expectations without providing the framework for learning to be collaborative and compounding rather than siloed. I’m looking forward to experimenting with corporate programs this year. https://sandboxtime.ai/corporate-events

What’s Next

I’ve been working with a coaching and consulting company called Talentism, whose methodology encourages learning by doing with rapid experimentation, which is exactly what Sandbox Time AI is built to invoke. The whole program is my own experiment in creating containers where people realize they’re capable of more than they think and leave inspired to push further.

If you’re telling yourself you’ll get serious about AI when things slow down, you’re the version of you that never makes it to the gym. Block the hour, show up, build something messy, wash it away, and come back tomorrow.

Six months ago, I connected with the HumanX organizers to help think through sessions, support speaker diversity and plan to host the customer experience roundtable. I built a WhatsApp group of 30+ people I knew who were going, picked out my sparkly sequin combat boots and a gold jacket so I’d be easy to find, and counted down the days. Walking into Moscone Center, I knew I was in for four full, exhausting, exhilarating days. I was right.

Hundreds of conversations across every domain: skills, tools, enterprise versus consumer, spatial versus language AI, near-term disruption versus long-term transformation. I occasionally bumped into the “nothing new here” skeptics but for anyone who showed up curious and open, there was extraordinary human intelligence, curiosity and learning in every room. Things are moving so fast that no one can track all of it, which is exactly why being in the room matters.

Here are my five biggest takeaways.

1. The gap between incremental and transformational

It’s easier to improve a step in a workflow than to rethink the entire thing. Faster loan approval versus a loan in five minutes. A better lead-to-demo handoff versus a reimagined buying journey. While the incremental gains are real, the “rethink what’s possible” transformation is where the competitive advantage lives. May Habib of Writer highlighted the importance that the people working on real transformation need permission to work across multiple domains and multiple steps. That requires executive sponsorship, a cross-functional agile team, and the mandate to start from scratch. Most organizations are doing the former while talking about the latter.

2. Your new CX competition isn’t who you think

I hosted a CX peer exchange at HumanX with 50 founders, CEOs, go-to-market, and CX leaders. The most striking observation of the session was that consumers arrive at your website or product with a new baseline expectation given their daily use of Claude or ChatGPT. They expect your organization to know them the way AI already does. That means your competition in customer experience is no longer just the players in your category. It’s the best AI your customers interact with every day. That’s a fundamentally different design challenge, and most companies haven’t caught up to the question yet.

3. AI native is the wrong label

The term “AI native” as a desirable hire implies age building on the “digital native” predecessor which assumes younger means more fluent. Interestingly, given most universities haven’t figured out how to balance training AI skills without compromising the critical thinking skills that working through hard problems on your own develop, plenty of college graduates are in their AI fluency infancy. What I observed at the conference is less a generational divide but rather one based on curiosity, learning and attention. The people pulling ahead: self-starters, experimenters, the ones blocking time to play regardless of age. Andrew Ng said it plainly: learn to code. Not a specific language but the fundamentals of how things get built.

4. Everyone is a builder now. Quality is everyone’s problem.

One of my favorite conversations was lunch with Ben Lingard, one of my favorite engineers I’ve worked with from my time at Guideline. He shared with the non-technical new builders at the table how he applies the traditional engineering process to agentic development: write up what you want to accomplish and then build and run agents for design sessions, implementation sessions, adversarial reviews, verification gates, PR code reviews for features and bug fixes. Ben’s insights combined with Annie Tsai’s hands-on AI agentic GTM expertise made for an awesome lunch table.

It’s easy to build today. For a standalone personal app, a prototype or a new user interface like a dashboard, it’s magical. However, it gets messy fast. The opportunity at this moment is to learn the fundamentals of how to build at scale.

5. The undercurrent everyone felt

The overall vibe at HumanX was positive and bullish. Founders with no legacy stack moving fast. Established companies like Zendesk and Intercom rethinking their entire business model with AI native offerings, disrupting themselves before someone else does it for them. My friend Tom Hale of Oura described the opportunity for healthcare shifting from monitoring sickness in a moment of time to a proactive, predictive system of managing long-term wellness. And for anyone who stayed curious beyond their own lane, Dr. Fei-Fei Li’s work at World Labs opened up the spatial AI conversation — robotics, healthcare, manufacturing — in ways that felt less like future-casting and more like now. My friend Lauren Dunfield CEO of Guidewheel, who is transforming factories with AI, could not agree more on the potential ahead.

Underneath all of it ran a quieter current: what does this mean for jobs and careers? Nobody had a clean answer. The optimistic frame of “growing businesses can do more with the same team” only stretches so far, and most people in that room knew it.

In summary

The conference was called Human[X] for a reason: human and AI capabilities are stronger together. Empathy, creativity, and complex judgment paired with AI’s speed, scale, and thoroughness.

However, even with the optimistic framing that growing businesses can do more with the same team, the honest reality is that most established organizations will become leaner. AI-native companies are already building next-generation workflows with very small teams who can build and manage AI agents.

What struck me was while HumanX organized itself into distinct tracks such as Ecosystem for investors, Customer Engine for go-to-market and Builders for product and engineering, in every roundtable, master class and hallway conversation, the lines have dissolved. Everyone is becoming a builder. Every function needs to understand go-to-market. Investors need to understand the full stack, from chips to applications. For someone who spent their entire career gravitating toward cross-functional, end-to-end thinking over staying in a single functional lane, this moment feels… right.

Where will I go from here? Stay curious. Block the time. Show up. Learn. Your journey of being AI fluent starts by taking the time to learn by doing. For me, it means hosting more SandboxTime.AI sessions.

In preparation for a CX Peer Exchange session I’m facilitating at the Human[X] AI conference in San Francisco, one question is top of mind: what conversation would make someone do something different immediately after the session? Here’s where I landed, and why it started with a photo I took in Japan.

The answer started with this photo taken at TeamLab Borderless in Tokyo. The experience is immersive, overwhelming, and completely unforgettable. Founder Toshiyuki Inoko didn’t start by asking what the art should look like. He asked what visitors should feel, and built everything backwards from that answer.

From One to Many

That’s the question I want to explore with my peers. I’m asking attendees to share three things about themselves and ask AI what that reveals about what they would expect as a customer and what it feels like when a customer experience is great for you. Do you feel smart? Supported? Confident? We will start personal and broaden to designing, measuring and monitoring CX at scale.

This week I tackled Ruben Hasid’s I am just a text file taste interview about my preferences, beliefs and feelings resulting in an 18-page profile I now hand Claude before any task. The insights made me wonder: what if the fastest path to design a great customer experience in the age of AI at scale is to start with the power AI has to unpack these expectations, perceptions and feelings from one customer? You.

The definition of customer experience is the cumulative, takeaway impression formed by a customer based on all interaction with your brand from marketing and sales to product and service. It represents the emotional, cognitive, and sensory response to these interactions. Capturing, understanding and acting on those insights at scale has totally changed with AI.

The Design for Delight Methodology Still Holds

My thinking about how to design for that feeling is rooted in the Design for Delight framework I learned at Intuit that I still believe in completely. Three steps:

1. Develop deep customer empathy. Fall in love with the problem, not the solution. The customer needs a hole in the wall. The drill is just how you get there.

2. Go broad to go narrow. Enumerate the possibilities before you narrow. Context is everything. Don’t rush to the answer before you’ve understood the space.

3. Run rapid experiments. Build the cheapest version that lets you learn something real. What used to be paper prototypes is now, frankly, mind-blowing.

The goal of all three steps is the same: understand what the customer needs to feel, and then design relentlessly toward that feeling.

When Data is No Longer the Barrier

For most of my career, too much data in too many places meant the insights were always out of reach. AI changed that. If the data barrier is gone, the hard part is now focus. Deciding what feeling you are designing for, and building everything backwards from that answer. Building has been democratized. Clarity is the competitive advantage.

Lenny Rachitsky’s recent AI productivity survey makes this concrete. Founders are getting dramatically more out of AI than most operators because they’re using it for vision, strategy, and design. Operators are more likely to use AI to complete tasks. In CX, that gap shows up as a conversation still dominated by efficiency metrics: deflection rates, resolution time, handle time. Those metrics matter. But they measure the transaction, not the feeling.

The more interesting question is what a great experience looks like when AI can hold the full context of a customer across every touchpoint. When recognizing patterns across mountains of data is no longer the hard part. Same tools. Different questions being asked.

Designing for delight is the same methodology it has always been. The speed is just different. The winners in this moment are the ones who move from confusion to clarity fastest, run the right experiments, and learn before anyone else does.

Better Together

What I’m most excited about is finding the builders and operators who are willing to say where they’re stuck. The ones asking real questions, running real experiments, and genuinely curious to learn what it means to design experiences that make people feel something worth telling a friend about in the age of AI.

This morning I was scared to paddle out but I did it anyway. I’m always a bit nervous surfing as my skills have never advanced past novice, and I’ve been extra cautious since a fin injury last year. I always check the ocean’s “energy” on Surfline, preferring under 500 kj. This morning it was 1,700 kj, well outside my comfort zone. How is this relevant to managing your career in the age of AI?

Point of View: Courage is a muscle. And like any muscle, you build it through practice, not perfection.

1. It’s the Journey, Not the Outcome

This morning wasn’t going to be measured by successfully catching waves. They were fast and steep. Success was having the courage to paddle out, stay in the water for 30 minutes, attempt some waves, and exit without injury.

With so many AI tools, features, and models being released, I’ve stopped measuring success by what I build. Instead, I measure it by time spent experimenting, trying, and playing to learn.

Try this: Check out my friend Kat Hill Contag’s app on Lovable AI recess. Spend 15 minutes with a challenge or playing with a prompt.

2. Community Makes All the Difference

I’ve had an amazing group of surf friends for a decade. When six friends said “in” to “Sat am surfers?” I joined. I’m the only woman, and most of these guys started surfing as teens. They’re much better than me. But there’s no judgment. We suit up together, enjoy the journey, watch each other’s backs, and encourage each other.

In the age of AI, create a community to hold yourself accountable. It’s a lot more fun.

Try this: Reach out to a coworker or friend to share what you are each doing to build in AI. Comment in LinkedIn on someone else’s AI experiment. Join an AI community (Discord, Slack, local meetup) and lurk if you need to, but show up.

3. Courage is a Muscle

I love that expression because it reminds you to practice doing hard and uncomfortable things. The mornings I surf, regardless of the outcome, I’m more likely to do things that are new and uncomfortable. I get out of my head and out of my own way because “If I can do that, I can do this.”

It’s easy to procrastinate on things that are new and uncomfortable. I don’t like being “inefficient” while learning new tools. I get nervous putting myself out there with a LinkedIn post. There are lots of reasons to avoid, procrastinate, stay in your head, stay in your comfortable space.

Try this: Do one uncomfortable AI thing this week that scares you a little. Automate a task you’ve been doing manually. Ask AI to review your writing and challenge you. Share something you’ve built. The discomfort is the point.

Closing:

When in doubt, paddle out. When uncomfortable, do it anyway.

Photos from this morning: Scouting from the cliffs of Half Moon Bay, walking out to the beach at Miramar, the waves after I got out.

What a Year Off Taught Me About Building in AI: That You Can Use Right Now

This is the Year of the Fire Horse. It comes once every 60 years and symbolizes passion, speed, transformation, and bold action. I’m not much for numerology or Chinese Zodiac, but the timing feels right. This past year gave me something most professionals don’t have: space and time to learn, think, and experiment with AI without the demands of a full-time operator role.

I completed executive coaching and board leadership programs. I built a capstone project in Stanford’s GenAI for Product Managers course. I deployed a full-stack application during Ember AI’s weekend bootcamp. I played with Lovable and tested Claude Cowork and Skills.

Here’s what struck me hardest. The classes and podcasts gave me a solid foundation. But the real transformation happened in the messy middle of debugging, deploying, troubleshooting broken prototypes and agents. What separates learning from doing is courage to try when you don’t have all the answers.

You don’t need a year off to get there. You need a different approach. Here are four frameworks from my year that work even with two hours a week.

Learn by doing, not just consuming

Students retain 75% of what they practice by doing versus 5% from lectures (National Training Laboratories). I watched this play out in real time. After hours of lectures, I learned more about building in AI in a weekend bootcamp playing in VS Code, restarting my local services, fixing broken flows, committing to GitHub, and deploying to production.

Start small. Check out Lenny’s Product Pass. Pick one tool this week. Build something tiny. A prompt that automates a weekly report. A simple chatbot that answers your team’s FAQs. Don’t aim for production-ready. Aim for functional. The outcome matters, but the learning matters more.

Learn together, not alone

I spent too much time in my house in 2025. Too comfortable. Too slow. I do my most creative work with others and with deadlines. Research from Carnegie Mellon shows that active learning with collaboration produces better outcomes than traditional lectures (Carnegie Mellon).

Find your people. Host a monthly roundtable where peers share 10-minute experiments: the problem, the tool, the lessons learned. Join a weekend bootcamp or workshop. Create accountability. Don’t wait for your company to provide structure. Build it yourself.

Join the race, don’t watch from the sidelines

I thought I might be done being a full-time operator. I thought I might consult, do fractional work, coach leaders. But this moment is too exciting to sit on the sidelines. This moment invites people who love to be on the race track. Who roll up their sleeves. Who operate independently. Who aren’t afraid to learn new technology. Who simplify the complex and bring order to chaos. This year, I’m joining an AI-native company to use my experience and energy to inspire others to be confident and do hard things.

You don’t have to change jobs. But commit. Stop dabbling. Pick one AI skill you’ll master this quarter. Ship something, even if it’s imperfect.

Be brave and bold in small ways

I’m writing this from Patagonia with my new travel guitar so I don’t fall behind in my intro class. I’m looking for a local dance class while continuing to Crossfit, bootcamp, surf, and hike. Bravery doesn’t require grand gestures.

What’s your version? Maybe it’s sharing an AI experiment with your team even though it’s rough. Maybe it’s volunteering to lead an AI pilot when you’re not the most technical person in the room. Maybe it’s asking the “dumb” question in a meeting because you know five others are wondering the same thing.

The fundamental transformation underway with AI aligns with the Year of the Fire Horse. This is not the time for comfort, overthinking, slowness, or timidity. This moment requires action, trying, being brave and bold.

You don’t need a year off to get started. You need two hours this week to build something. To find one person to learn with. To commit to one skill. To take one brave step.

The Fire Horse year favors those who choose courage over comfort, action over waiting. I invite you to embrace it with me. Learn by doing. Learn together. Commit. Be brave and bold.

Over the past six months, I’ve joined thousands of people experimenting with vibe-coding platforms to turn ideas into applications. The magic fades the moment you move beyond a simple landing page. Databases, authentication, API keys, and integrations quickly turn inspiration into error messages especially if you don’t come from an engineering background.

Like many AI build hobbyists, I’ve wanted access to an AI-savvy engineer who could provide just-in-time guidance so I could learn faster and get my growing pile of half-built prototypes to production. For those non-technical professionals trying to translate their organization’s complex internal workflows into AI-powered tools, the challenge is even more daunting.

Enter Ember AI, which pairs a vibe-coding prototyping platform with real engineer support to help non-technical professionals build end-to-end applications. This past weekend, I participated in Ember AI’s Build Bootcamp. Here are six highlights and lessons learned from my MLK weekend.

Think and build in chunks

My first draft of my application idea was a simple one-pager. Using Ember AI’s Idea to MVP chat interface, I turned that into a minimally viable PRD, something realistic to build and deploy in a single weekend.

I asked ChatGPT to find gaps and weaknesses, which helped me think through features, user stories, failure and recovery states, branding guidelines, and more. My mistake was dumping that six-page brainstorming document into Ember AI’s PRD Generator for an MVP build as it created confusion instead of clarity.

My next PRD will be intentionally high-level, succinct and focus on the core problem I’m trying to address. I’ll refine and build on my idea section by section with LLM support rather than front-loading too many details at once.

Roll up your sleeves and open the hood

In 1994, I learned Unix and HTML to understand how the internet worked.
In 2014, I led a data-center migration to understand how servers, networks, storage, databases, virtualization, applications, and services fit together as we shifted into cloud hosted environments.
Today, I’m building in AI to understand how vibe-coding platforms, LLMs, VS Code, GitHub, MongoDB Atlas, and API integrations actually work.

Each time, the pattern is the same. Understanding the technology stack behind the scenes makes me a better business partner and leader. It’s humbling and energizing.

Learn by doing, even when it’s uncomfortable

After hours of debugging with Haziq, my engineering partner, walking step by step across multiple tools, I secretly wished he would just log into my account and fix the issue for me. I just wanted a green “live” message rather than another frustrating red “failed” error with log-file analysis in VS Code.

When I got my app to work with Haziq, I was so excited that I didn’t care that there was a broken flow with missing elements. I figured it was good enough. But this morning on my own, I opened VS Code, restarted my local front- and back-end services, prompted a fix for the missing page, tested locally, committed and pushed to GitHub, rebuilt the public site, and it worked. The outcome mattered, but the learning mattered more.😊

Trash the code, keep the learning

In Stanford’s Generative AI for Product Managers class, we talked a lot about building and testing with the goal of throwing your work away to carry the learning forward to the next build. This is a mindset shift from the historically heavy lift of planning & prioritizing before a single line of code is written to protect scarce engineering resources. Today, writing code is cheap. Learning how to use AI effectively is valuable. Build quickly with the expectation that what you’re really shipping is insight for the next iteration.

Test early and often

While building is “cheap,” testing and debugging are still “expensive.” After my first build, there were plenty of issues: ugly branding, broken flows, missing pages. With Haziq’s guidance, I tackled them one at a time by prompting a change, testing locally, committing to GitHub, and deploying publicly.

Any engineer reading this is probably thinking, no kidding - that’s how software engineering works. But for non-technical folks, this discipline can feel uncomfortable. In our usual work, a document or decision that is missing context doesn’t fail loudly with error messages and generated logs providing breadcrumbs to address the root cause. Learning to build software is humbling, frustrating and empowering.

Ask for a better way

Building in Ember AI reinforced something we discussed often in the GenAI PM course which is the power of using LLMs as thought partners. Prompts like:

• Is there a better way?

• What am I missing?

• How could this be simpler?

• What are the alternatives?

• Show me your reasoning.

AI helps counter a very human tendency: falling in love with a solution rather than the problem and spending all your time on a happy path rather than eliminating friction. Building in AI makes it so much easier to consider risks, edge cases, and failure modes, and ultimately to stay grounded in the customer problem you’re trying to address.

This moment in technology is incredibly empowering and exciting. However, it requires curiosity, humility, creativity, persistence and patience to take advantage of that power. No amount of time in class discussions, reading LinkedIn posts and Substack articles, listening to podcasts and watching YouTube videos can replace the learnings from my weekend building in Ember AI.

While the platform created the code, the humans created the learnings:

My gratitude to the humans

• Ha Nguyen: An afternoon coffee conversation connected me to the Qasim and the Ember AI team. You are the ultimate connector and I can’t wait to build with you.

• Qasim Salam: Your journey building the Remotebase and Talently platforms to find, interview and train engineers unlocked the ability to found Ember AI with amazing engineering talent.

• Garrett Kelly: You are a master facilitator, teacher and operator. Your journey building in Ember AI is inspiring and your encouragement that we can do this inspired the entire bootcamp.

• Haziq Ali: You believed in my ability to build and deploy an app more than I did. You patiently walked me step by step through new tools, products and processes and celebrate “go live”

• Daniel LeSieur: You were a fantastic mentor providing us newbies with practical tips and tricks, support and encouragement.

• Deb Liu: Our paths have “almost crossed” for many years but in addition to encouraging your Perspective Substack, I’m stoked to support your journey with Ember AI.

As AI makes creation faster and cheaper, the differentiator shifts from who can write code to who can learn effectively. Leaders who understand how these systems work will ask better questions, make better decisions, build stronger products, and partner more effectively with technical customers and team members. This weekend reinforced for me that the future belongs not just to builders, but to learners who are willing to build.

My live application: GiddyUp

I intended to build an application that lets users easily host multiple events to a regular friend group. Future functionality would include streamlined chat for logistics, Doodle-like scheduling for finding the best date, calendaring, email and text invites, sub-groups of friends for event series and ability for friends to host their own events.

What I actually built is a very simple event and friend group creator with OpenAI GPT-4 Mini in the background with basic account creation and a dashboard view. The outcome is simple but it works and I learned a ton. Check out GiddyUp

Last year, I wrapped up my most recent full-time operator role at the same time my kids headed to university and my husband pursued his lifelong dream to be a full-time commercial pilot. As a result, I have space in my calendar for the first time in thirty years to learn, explore, listen, think, tinker and connect at this pivotal moment of AI transformation. This moment reminds me so much of 1994.

In 1994, we were in an economic recession with almost no hiring. Desperate to land an entry-level job, I was intrigued by a San Francisco Chronicle article that claimed that there was one exception to the lazy, useless GenX grunge generation and his name was Marc Andreessen. Also a recent college graduate, Marc had just built this technology called Mosaic which was a way to view or browse this network of computer in academia called the World Wide Web. The next day, I bought “Unix for Dummies” so I could create a Unix based email account through the California Research Lab to learn more.

The following week, I snuck into a small Saturday morning Unix developer meetup at Pacbell’s San Francisco’s headquarters to meet this fellow 23 year old that not only had a job but had started a company called Netscape which was not a thing then. As I watched as men twice his age hang on Marc’s every word, I knew something bigger was at play.

When Marc sat down next to me after the Q&A, I whispered “Does this attention freak you out” at which he replied “You have no idea how weird things have gotten”. At the time, Netscape was a thirty person company above an ice cream shop in Mountain View getting ready to launch Netscape Navigator. I still laugh thinking back to that moment.

In the weeks that followed, I was laser focused at having a front row seat in this new world. I landed a data-entry job for $19.5K/year at a young company called Macromedia so that I could be “in the room where it happened”. After months learning HTML and helping build our first web site after hours, I became the second person on the new web development team. In 1996 I built Macromedia’s e-commerce system in Netscape’s Merchant System and in 1997 I landed the product manager role for Shockwave and Flash which was the first animation on the internet. Thank you Bud Colligan, Joe Dunn and Norm Meyrowitz for believing in me.

While thirty years separate the Internet’s start in the mid 1990s and GenAI explosion today, both moments require time and space to develop a point of view of what it means for you and your career. If you are in a full-time operator role, your focus is likely what AI means for you in your current role. If you are a leader, you are trying to navigate what it means for your team, your function and your organization. Creating time when you are already scheduled in thirty minute increments is a challenge.

There are a some with the financial means to step out of their full-time operator role to invest in finding their own front row seat in this GenAI revolution. However, for those trying to explore what’s next while balancing the demands of your full-time operator role, consider these 7 practical ways to become a more savvy AI builder.

In those early days of the Internet, there was no YouTube to watch, Github to reference or Reddit to read. Learning was with books like “Unix for Dummies” or Ralph Kimball’s “Data Warehouse Toolkit” and then putting in the hours to figure it out. To hold yourself accountable, I invite you to create an end of 2025 AI learning goal. Ping me if you’re interested in partnering to do so. As for myself, I’m off to meet up with a fellow classmate from my GenAI for Product Managers course to submit an application for Lovable’s SheBuilds buildathon which is due tomorrow.

Stephen Covey highlighted human nature’s bias to prioritize the urgent over the important. We are prone to respond to Slack and prepare for our next meetings while punting critical career development work to late nights and cross-country flights. Here are seven practical ways I’ve improved my AI building skills in 2025.

1. TAKE A CLASS: AI is a technology that’s accessible to all so doesn’t necessarily require a class to learn. However, taking a class can be a great way to block your time and hold yourself accountable. There are lots of on demand classes options for those disciplined enough to prioritize the time each week. However, if you are like me and workout in group classes to drive accountability, sign up for a live class with a hands-on vibe coding project. I have loved Vikash Rungta’s GenAI for Product Manager class at Stanford this Fall as it has been a fantastic AI foundation with an engaging community and a vibe-coding group project. Check universities, top tech companies or learning platforms like Coursera for in person or online live courses.

2. BOOTCAMP OR WORKSHOP: To help leaders up their AI skills, there are bootcamps evolving where you learn to build and publish your native AI application. SnapCamp is a three week vibe coding camp to go from idea to product with AI tools, engineering support and no coding experience required. If your full-time role prevents an intensive camp, consider a hands-on workshop at an upcoming team offsite with a cadence for follow up and accountability. Dan Olsen has developed an in person vibe coding for product teams. I’m working on a workshop tailored for customer, operations and product marketing teams. Learn more here or direct message me if you’re interested.

3. HOST A ROUNDTABLE: If you really want to learn something, teach it. Host a session for professional peers to share learnings from each others’ AI experiments. For the CX AI Leader Roundtable, I invited twelve peers for monthly sessions. We rotate leaders to share their insights in ten minutes with three slides: the problem you were trying to solve, the experiment you ran (what, who, why, how, when) and the lessons you learned. See Insights from the CX AI Leader Roundtable.

4. LINKEDIN: My LinkedIn algorithm now leans heavily into AI enthusiasts and AI native builders for snack-sized insights, trends and advice. I’m particularly drawn to leaders who ask followers thought-provoking questions such as Hubspot CEO Yamini Rangan and Gamma CEO Grant Lee. A few specific posts I found insightful include Donna McCurley’s take on GTM, Jules Boiteux’s journey of vibe coding and Sharmin Ali’s perspective on a 3-member AI company.

5. NEWSLETTER & PRODUCT PASS: While there are many great AI newsletters, there’s a reason Lenny’s Newsletter breaks through the noise with 1.1M subscribers on Substack. In addition to great content, Lenny has a $350 paid subscription that includes a ProductPass to leading AI native tools such as Granola, Lovable, Replit, ChatPRD, Gamma and others. Even if you don’t want to publish an application, it’s important to play with multiple AI native applications to better understand design, development and customer onboarding.

6. PODCASTS: I’m a fan of Matt Turck’s Machine Learning, AI and Data or MAD series. Matt just released the 2025 MAD Landscape report with the 25 trends and an interactive map. I listen to the MAD podcast regularly. Some recent favorite episodes include Nathan Benaich’s “State of AI 2025” and Granola CEO Chrise Pedregal “How to build a Beloved AI Product”. I’m also a fan of Claire Vo’s How I AI for practical ways to improve the quality and efficiency of your work.

7. CONFERENCES: There is value in attending vendor conferences like Dreamforce and Inbound particularly if you are a hands-on with their technology so you can appreciate their AI capabilities. Consider attending an AI specific conferences such as Human(X) in San Francisco in April 2026 where I’m excited to host a roundtable. Sign up at HumanX.

When you share your goals to another person, your chance of success rises from 4% to 65% (source). I invite you to share in the comments one thing you will do to invest in your AI building skills in 2025. Posting four AI articles in November to help leaders navigate the AI transformation is my goal. And you?

In the Spring of 2020, I tried to figure out how to lead my customers’ “pandemic” experience alone in my home office. Rather than navigating the uncertainty isolated, I gathered a dozen CCO/CX leader friends on Zoom to talk through our challenges, actions and learnings.

What started as a single meetup with a dozen professional peers evolved into a monthly recurring session sharing perspectives and advice on growth, cross-functional partnerships and measuring success. We called it the “CX Leader Roundtable”.

In 2025, GenAI’s disruption to the tech workforce has raised many of the same feelings of isolation and overwhelm. Leaders struggle to keep their teams engaged, customers satisfied and shareholders delighted while simultaneously transforming their function and business with GenAI. This needs to be done with the same or reduced teams since the productivity benefits from AI are expected the moment the initiatives are created.

We re-launched the “CX AI Leader Roundtable” as a confidential, non-vendor sponsored space to discuss our AI transformation journeys. We are testing a new framework where each participant shares a three slide, 10 minute TEDTalk style AI initiative with problem statement, experiment, lessons learned and next steps. In addition to learning with and from each other, there is the benefit of having a forcing function to synthesize your impact outside of your organization.

The first topic was how one customer success executive launched a “Renewals AI Agent” that CSMs could use to navigate customer renewal questions. Built from two years of recordings from the internal renewal executive meeting combined with company policy guardrails, the AI agent was piloted, deployed, refined and scaled to help CSMs get the guidance they need to help customers.

Our second session focused on learnings from building, deploying and scaling customer-facing GenAI agents for customer support. Both customer support executives shared similar learnings and recommendations:

• Start with low-complexity, high volume well documented cases to assure customers get meaningful answers from day one

• Invest in a content person or team whose day job is to clean up and maintain your customer facing help center

• Hire more experienced or domain specific customer support talent as the low-complexity entry level work is covered with AI

There was a great discussion in the ongoing challenge of measuring the impact of AI. Financial metrics matter most so flat or reduced headcount cost with the same or rising customer cases is the bottom line measurement particularly with your CFO. Most leaders agreed that while deflection is mainstream it’s not particularly meaningful given the challenge of estimating what didn’t happen. We explored the concept of measuring the percentage of “containment” which means that a customer’s questions didn’t require human followup because the AI agent resolved the issue completely.

Almost every leader I meet in a SaaS tech organization seems to be struggling with the daily demands of their functional jobs while trying to keep up with the dizzying pace of AI innovations and transformation of their organization.

While there’s no silver bullet, carving out time each month to meet with and learn from your peers is a way to address the challenge as a community. The forcing function of time set aside each month to learn with your colleagues and being accountable to contribute can be just the right level of accountability. #BetterTogether

When I commuted to Intuit years ago, my favorite program was NPR’s “This I believe”. Listeners submitted essays about the core value that guide their daily lives. Selected essays on topics ranging from acts of kindness, identity and patriotism are read by the author. Every time I listened to an episode, I would spend the rest of the drive trying to land on my own “This I believe”.

Fast forward to today when I’m working with a high school senior through the college application process as a mentor at the Boys and Girls Club. My mentee has to share a perspective on her leadership, creativity, and talents articulating how she will contribute to her future community.

Since wrapping up my last full-time leadership role, I’ve had time to think about I can leverage my cross-functional leadership experience in tech to add the most value to my professional community. This opportunity for reflection happens to coincide with the fundamental shift in how businesses will transform their work with the arrival and adoption of GenAI.

I have so much empathy for my colleagues in full-time leadership roles trying to keep up with the pace of GenAI’s impact, while driving transformation on top of their daily demands of running their business. Almost everyone I meet with is exhausted and overwhelmed.

My objective with this new AI Angles newsletter is to share perspectives on how GenAI is impacting leadership, organizational design, and customer experience based on my experience as a former executive in tech and a current builder in AI. My goal is to offer “This I believe” perspectives to help my network of overstretched leaders in tech navigate the AI transformation with a bit more support and confidence. I invite feedback, questions and your perspective as one this I truly believe is that we are always #BetterTogether.

1. AI expertise on boards Just as financial and cybersecurity skills are now table stakes, boards need members with technical transformation, data, and AI expertise. Yet 66% of boards have little to no AI knowledge, and 40% are rethinking composition (Harvard Law School Forum for Corporate Governance.

2. Cloud, clean, and curated data AI requires more than curiosity and budget. Enterprises need clean, curated data and documented workflows in the cloud as the foundation to scale. Without it, AI transformation stalls before it starts.

3. Innovation vs. incrementalism Every transformation needs a clear north star: What are we solving for, and how

will we measure success? If AI goals are limited to efficiency and automation, organizations risk missing the real opportunity: innovation. Boards should ask: How would an AI-native startup disrupt us? Customer-centric, “design for delight” playbooks show the way.

4. Skills over functions AI demands agility, not rigid functional silos. Because AI shifts from deterministic to probabilistic outcomes, teams need new strengths: experimentation, analysis, iterative planning, documentation, communication—and the courage to test and adapt.

👉 With 95% of organizations reporting no ROI from generative AI (MIT report), boards and CEOs will be asking why—and what needs to change.

The takeaway: oversight alone isn’t enough. Boards and C-suites need the same agility, curiosity and agility skills to lead meaningful transformation and be the one to disrupt themselves.

Subscribe now

I recently completed Stanford University Leading in the Age of AI course and walked away with four core takeaways that apply to every executive navigating this moment. Whether you’re experimenting with AI or leading full-scale transformation, these insights can help you move faster with purpose.

1. Leverage Technology Transformation Playbooks

Many companies start their AI journey by enabling features in tools they already use such as Salesforce, Hubspot, Intercom, etc. It’s a smart first step: functional leaders can run controlled experiments without needing full organizational buy-in.

But true innovation, whether reimagining internal workflows or launching AI-powered customer experience, requires more than incremental experimentation. It’s a company-wide transformation.

While GenAI is moving faster than the internet in the 1990s or cloud in the 2010s, this course reminded us that many principles from past tech revolutions still apply. Leading transformation demands strategic clarity, executive sponsorship, and cross-functional collaboration. The best leaders treat this like any other major transformation effort with urgency, vision, and structure.

2. Drive Relentless Strategic Clarity

During my decade at Intuit, I had the privilege of working under Steve Bennett and Brad D. Smith, leaders who made strategic clarity a team sport. To this day, many of us refer to it as a master class in aligned leadership.

That same clarity is critical when leading in the age of AI. The fundamentals still apply which means start by answering these six questions:

• Why: What problem are we solving? How will we measure success? What’s the unifying goal across teams? Tip: It’s often a measurable customer, employee, or shareholder outcome.

• What: What initiatives will get us there? What are the three things that matter most?

• Who: What cross-functional skills are required? Who brings those skills?

• When: What key milestones such a a leadership offsite, company all hands or board update can serve as forcing functions? What’s the story we want to tell in six months? Draft that slide now.

• How: How will we drive accountability and create a culture of experimentation and learning?

• Where: Where’s the single source of truth? Hint: It’s not just a Slack channel.

3. Be Purposeful with Team Culture

At Intuit, I participated in a leadership program called Fast Path where high-potential leaders collaborated on a real-world project. My team explored what made certain cross-functional teams outperform and why others stalled.

The key insight: the “heart” of the cross-functional team must be stronger than the pull of the functional “corners.” The teams that succeeded had a shared purpose, strong trust, and clearly defined operating mechanisms that fostered open communication, fast learning, and joint problem-solving.

That culture is even more critical in AI transformation. GenAI is inherently experimental: you learn by doing, testing, prompting, refining, and sharing. It thrives in cultures that value transparency, bravery, and cross-functional trust. My favorite hashtags apply more than ever: #OneTeam #BetterTogether

• A short exercise to create a high performing team fast

• A winning cross-functional team’s secret sauce

4. Prioritize Skills, Not Just Functions

At SurveyMonkey, I led Customer Experience while also supporting a global community of CX leaders using our tools to drive impact. In that “meta” role, I observed which teams could turn insights into results and which couldn’t. The best teams focused less on job titles and more on skills. The same holds true in AI transformation. Some key skills include:

• Cross-functional connector who aligns teams and facilitates momentum

• Data & analytics expertise to translate insights into decisions

• Business systems acumen to unify and leverage data

Like CX leadership, AI transformation isn’t about installing a tool. It’s about building a feedback loop: learn, prioritize, act, refine. That mindset combined with the right mix of skills makes all the difference.

• Three skills to be an effective CX Leader

Being back in a classroom on a college campus was energizing especially with a cohort of leaders from different industries, functions, and generations. We’re in the early days of this transformation, and I couldn’t be more excited to coach, mentor, and advise high-performing leaders on this journey. Let’s leverage the best of what we’ve learned from past tech revolutions to win in the age of AI.

• Learn more at christinerimer.co

  m

At the NACD (National Association of Corporate Directors) AI Technology & Innovation Master Class, four insights stood out with each sparking important questions about how boards and executive leaders can better prepare for AI-driven transformation. I’d love to hear what resonates most with you.

1. AI Strategy Prioritization & Pace

Over the past three years, companies have been under pressure to operate leaner and burn capital more slowly. While this discipline is necessary, it leaves senior leaders increasingly stretched balancing day-to-day responsibilities such as leading teams, serving customers, building products, driving growth while also trying to “hold space” for transformation through AI.

If the speed at which your organization executes an AI strategy is a competitive differentiator, then how can the board and executive team ensure they’re truly making space for it? Is AI a line item on the strategy agenda, or is it buried under operational noise?

2. Board Structure & Ownership

Where does AI oversight belong in a board’s committee structure? The conversation revealed how complex and cross-functional AI governance really is:

• Audit Committee: Compliance and financial reporting implications

• Compensation Committee: Incentivizing transformation through rewards

• Nominating & Governance Committee: Ethical and reputational guardrails

• Risk Committee: Core business risks in AI-driven workflow changes

But beyond defense, AI requires offense. Boards should ask: do we have a Strategy or Planning Committee explicitly focused on how AI can unlock value and drive long-term growth? One striking metaphor stood out: An AI strategy can’t be treated like disaster recovery or succession planning with an annual presentation that’s shelved the rest of the year.

3. Board Member Experience

Selecting the right board member composition is always about strategic alignment with the company’s stage, market, and risk profile. Typically, boards prioritize expertise in:

• Finance & Audit

• Technology & Cybersecurity

• CEO Operator

Other functional skills like go-to-market, product, legal, or international experience vary by company needs. Given the rapid evolution of AI, it may be unrealistic to expect current board members to be deep AI experts. However, directors who have led organizations through prior tech transformations like the internet in the 1990s or cloud computing in the 2000s bring valuable operator experience to the table.

4. Questions to Answer

One of the best phrases for a board’s role is: “Nose in, hands off.” Engaged directors govern well by asking smart, forward-looking questions. Two stood out:

1. How would you disrupt yourself? There is value in stepping into the shoes of an AI-native, agile competitor and map out how they could take you down. Where would they start? Biggest opportunity? How does this change where you focus?

2. Who is in charge of your AI strategy? AI impacts internal workflows, productivity, and customer experience. So who truly owns the AI strategy? While C-suite sponsorship is essential, a leader’s functional orientation can create bias. Boards should push for clear, cross-functional leadership of AI initiatives.

Thank you to the many insightful speakers including Nora M. Denzel, Amy Wilkinson, Ryan McManus, and Roger Roberts for sharing provocative and practical perspectives. Join the conversation with thoughts and perspective in the comments.