Same Same. Different Different: Product Leadership in the Age of AI

Same Same, Different Different

In January this year (2026), I was asked to speak on Product Management and AI at a brown-bag lunch. So, I wrote a new talk outlining my observations over the last six months in product. As I sit behind my keyboard, I see that much appears to have changed in Product globally, but if we peel back the layers, the core principles of product still hold and the question becomes one of product judgment and who is responsible for building it.

Perhaps as a reaction, I see that people are starting to distinguish between an AI Product Manager and a Product Manager. To be honest, this pattern followed the same pattern when Agile was introduced. Many folks became an Agile Product Manager.

So, I titled my talk “Same Same & Different Different”, the fundamentals haven’t changed, but the environment and the execution demands are different. And to be fair, we may need additional capabilities in our toolkit because of the change in environment.

So, my talk in a nutshell:

What’s Same Same:

• Clarity on who our customers are and which problems are worth solving
• The courage to make real trade-offs, grounded in vision and strategy
• Commercial discipline. Delivering the returns the product promises, not just momentum
• Foresight to defend today’s position and create tomorrow’s options
• Judgement to interpret data in context and decide where to go and why
• Accountability. We own the decisions that drive outcomes.

What’s Different Different:

• AI removes the friction around efficiency and enables us to do more, faster
• The number of viable options for us to pursue is exploding
• Faster cycles mean market feedback comes sooner
• Engineers and designers increasingly operate as independent product creators
• We’re no longer only designing for humans. AI agents are becoming part of the user ecosystem
• New competitors don’t always look like companies; sometimes they’re individuals solving very narrow problems extremely well

AI Exposes Poor Product Judgment

The underlying reality, though? AI exposes poor product judgment.

When I asked the product community for feedback before delivering the talk, I learned a new term: premature certainty. It’s coined by Michael Fountain, and it perfectly captures the trap product leaders are falling into right now.

We have access to AI tools that generate new ideas with a swipe of a prompt. Executives love throughput. But more ideas without judgement is just motion. And we’re seeing the consequences: feature bloat, diluted focus, teams chasing “false validation” instead of building conviction.

This is the moment where judgment matters more than it ever has.

Judgement doesn’t come from AI-generated analysis. It comes from experience, particularly experience with failure and how we recover from it. It comes from immersing yourself in organisational and cultural context, something AI can’t do unless you feed it a ton of data; it still won’t fully understand. It comes from watching your colleagues and reading their reactions to what’s being presented. It comes from synthesising all those elements and being brave enough to make a decision, even when you know it might not be right.

Product leaders are operating in a climate where expectations are higher and tolerance for uncertainty is lower. AI is accelerating change, but not clarity. Trust, confidence, and decision-making matter as much as frameworks and processes.

So, in the era of AI, product leaders need to strengthen their ability to make good judgment calls. More importantly, we need to teach our product people how to do the same.

Building Product Judgment in Your Team

Building judgment in your teams requires four disciplines:

1. Make your thinking visible.
   When you decide, explain your assumptions out loud. Say what you expect to happen, and why. Then, when results come in, revisit it. Show them how you adjust when your assumptions are wrong. Judgment isn’t about being right; it’s about learning fast when you’re not.

2. Give them decisions with guardrails.
   Don’t just ask them to decide. Ask them to justify their decision.But be specific: What assumptions are you making? What would have to be true for this to work? What’s the first signal that you’re wrong? Is AI helping or hiding your uncertainty?

   This forces them to articulate their reasoning, not just their conclusion.

3. Normalise bad calls.
   Don’t just avoid punishing mistakes. Try to “celebrate” the ones people learn from. Yes, this is difficult. If someone made a decision that didn’t work but can clearly articulate why it didn’t and what they’d do differently, that’s a win. If you punish wrong decisions, your team will stop making decisions at all. They’ll just wait for AI to give them cover.

4. Set decision-making boundaries.
   Be clear about which decisions require judgment and which don’t. Not everything needs deep thinking. If you treat every decision as equally important, your team will burn out or default to analysis paralysis. Teach them to distinguish between:
   • Reversible decisions (make them fast, learn as you go)
   • Irreversible decisions (slow down, get it right)
   • No-decision-needed (this is noise, ignore it)

If we don’t coach, if we don’t create these guardrails and boundaries, our team will be rudderless because the tools will keep multiplying. Its capabilities will keep improving.

The question is whether we as leaders can build the discipline our teams need and have the courage to enforce it.