Leading through the transition to AI
Amidst all the excitement, the world for anyone who works in tech is filled with uncertainty. There's a paradox to this moment.
On one hand, we're living through one of the most exciting breakthroughs in the history of technology. (Sundar Pichai called AI “more profound” than the advent of…fire.) For the first time, you can go from an idea to a working prototype in hours. You can analyze data without writing SQL. You can generate designs without years of design training. The agency to build stuff is unprecedented and energizing. And the tools are getting better fast.
On the other hand, that same reality creates real uncertainty. When the capabilities expand this quickly, it's natural to wonder what it means for how teams are structured, what skills matter most, and where we fit into all of it. Anyone paying attention feels some version of this tension.
That leads to the second topic I want to raise. Amidst the many posts about our changing world, I’ve found most lack an answer to the most important question: What should we do about it? So, this is how I’m thinking about it.
What we know (& don't)
A few things we can assert with confidence:
AI is not hype. The capabilities are real and improving fast. It’s not just about learning to work with the tools we have today, but learning to evolve as the ecosystem progresses rapidly.
The boundaries between functions are shrinking. AI is blurring the lines separating roles as it gives everyone the power to be a generalist.
The nature of work is shifting. Greater efficiency changes the equation on team structure, skill mix, and how we spend our time.
Speed is a core tenant of the tech industry. Where some industries stall in inertia, tech’s nature is to embrace change. Most of this is good. Some of it is FOMO.
What we don't know, though, is…most of it.
We don't know what the optimal product team looks like. We don't know exactly what it means to be an engineer, designer, or PM with AI at our fingertips. We don't know how the industry (...or society) will respond. The sheer volume of unknowns is, itself, the most important thing to internalize. Certainty is not arriving anytime soon. If you're waiting for someone to hand you a playbook, you'll be waiting a long time.
What to do about it
Uncertainty is not an excuse for paralysis. Here's how I'm thinking about channeling this productively.
Get deep in the tools, and keep going. This is the single most important thing you can do. But don't fixate on mastering any single tool or workflow, since they'll keep changing. Learning how to learn new tools is the point. Expose yourself to new ones constantly. Build a habit of experimenting even when there's no immediate need, because the pattern recognition you develop across tools is what lets you adapt when things shift.
Embrace self-sufficiency and accountability. No matter your role, the expectation will be that you operate independently. Learn to conceive, analyze, design, and build stuff on your own. Roll up your sleeves. For simple things, e2e development will be autonomous. For complex work, you should drive prototyping with the assistance of specialist talent to harden the implementation.
Double down on PM fundamentals; they matter more than ever. I wrote previously about how AI will evolve the PM role. As AI compresses execution, the premium shifts to judgment.. Building gets easier. Knowing what to build does not. Here's what that means in practice:
Spend more time in the problem space. When execution is cheaper, the cost of solving the wrong problem goes up. Pressure-test whether something is worth doing before you optimize how. Sometimes easier execution just reveals that the idea was marginal to begin with.
Stay close to your users. AI can synthesize research and summarize feedback. It cannot replicate genuine empathy. Commit to a regular cadence of direct user contact, like calls, directly attending UXR, etc. Don't let AI become a layer of abstraction between you and the humans you're building for.
Cultivate taste. When anyone can generate output, curation matters more. Knowing good from great (and having the conviction to push for great) is a differentiator. Build it through reps: consume great work, study what makes products resonate, develop strong opinions about quality. The privilege of building fast is not license to ship AI slop or make users suffer poor product thinking.
Lead others through this. If you manage people or influence a team (read: all PMs), this one's for you. Your team is watching how you respond to this moment. If you're visibly anxious and frozen, they will be too. If you're curious, experimenting, and engaging openly with the uncertainty, your team will mirror it. Create space for your team to learn and experiment. Normalize the discomfort. Share what you're trying and what you're learning, including what isn't working. The leaders who will define this next chapter aren't the ones who pretend to have all the answers. They're the ones who demonstrate that you can move forward without them.
Anti-patterns
Stuff to avoid:
Don't bury your head…or catastrophize. I’ve seen lots of people suggest, “Every role is changing, except for mine.” (Marc Andreesen referred to this as a standoff between PM, eng, and design.) Pretending nothing is changing is not a strategy. The shift is real and it's accelerating. Conversely, spiraling based on speculation doesn’t help. If you’re worried, channel that energy into building skills.
Don't wait for AI playbooks. No one is going to hand you a curriculum or a company-mandated framework for how to adapt to AI. Just dive in, talk to colleagues, and figure out what works for you.
Change is a constant (but it's different this time)
Across my career, I’ve seen platform shifts (web to mobile) and the hype cycle play out with crypto and the metaverse (and many other mini ones…remember when Clubhouse/audio was the future of social networking). I’ve seen the industry shift in the way we work, like the initial transition to remote, the Great Resignation, and the power dynamic shift from employee to employer. Even so, I’d argue the pace of change and breadth of impact to the way we work is not comparable to previous cycles.
So, the people who will define the next era of product development are the ones running at the problem & opportunity. The ground is shifting. You can either let it shift beneath you, or you can learn to move with it.