Product Principles, Revisited

How is AI changing the fundamentals of product practice?

Three years ago, I created seven guiding principles for the work of Kin + Carta’s Product Practice (now part of Valtech). Designed to support brilliant work, they served us well – anchoring work across multiple industries, technical contexts and opportunity types. I read them as reminders daily.

However, as AI impacts the way we work, it feels like the perfect time to step back and ask, ‘Do these product principles still hold true?’

Outcomes Over Outputs

“Define success as the impact you make, not how much code you deploy. Focus on making things better in tangible ways.”

The fundamental principle of product management. That was true three years ago and it’s true today. Why? Because there’s nothing strategic about driving in the wrong direction really, really fast.

But what if movement drives insight and understanding? AI is radically changing the cost of shipping: this fundamentally changes risk profiles throughout the product development lifecycle and the way we work.

So, while the principle holds true, success now demands human and AI alignment towards clear outcomes, placed in the right context.

Problems First, Solutions Second

“Building things nobody wants is pointless. Think deeply, frame and distil the problems you wish to solve. Powerful ideas will spring from this insight.”

AKA: Don’t build solutions in search of a problem. But I do wonder if the linear framing of this principle is now coming under strain. It’s increasingly easy to (re)build and learn on the fly because ‘learnings’ (never mistakes) are less costly. This starts to devalue traditional, up-front discoveries.

But… by the same logic, continuous discovery becomes more valuable, not less. Staying close to users and the data, responding effectively to what you learn – and then going again. From this perspective, AI-powered product management will make us more agile and increasingly attuned to user and business needs.

Make it Simple

“The best products feel natural. The best processes are lean. The best solutions are easy to explain. Advance beyond complexity and ambiguity, it benefits everyone.”

The principle is about clarity of thought and execution, because strategy is, famously, about saying no to a thousand things. Good design (whether product, service, process or organisational) is about removing clutter and waste to deliver impact. This hasn’t changed.

Like caffeine, AI lets us do stupid things quickly. If we simply augment existing work with AI, this may lead to feature pile-up and process bloat.

But now is the time to rethink workflows and interactions. If we step back and re-focus on why something exists, perhaps the familiar can be radically transformed, replaced or removed. So ask, how can this interface be more elegant? How might we reduce handoffs and waste in this system? Which rote tasks can be removed, to leave more time to innovate?

This principle hasn’t moved, but the opportunity it points at has grown.

Work Together

“Combine viewpoints and talent from different fields. Give them a shared goal and help them solve problems. True teams build the best products.”

Today’s popular narrative is about solo makers. I’ve been doing some of this myself, building a Product Sense Coach and other tools with Claude Code, and it’s a lot of fun to do. But I believe this principle still stands. Because it’s about challenge, not headcount. Those moments where different perspectives and abilities come together are often where the magic happens.

AI doesn’t remove the need for any of that.

And yet… I increasingly instruct models to challenge my ideas. Perhaps I just need something to argue with, but I believe the need to work together remains, no matter how we think about ‘the team’.

Get Feedback

“Users and colleagues are a great resource. Speak to them. Test ideas with them. Release early and analyse performance. Create fast feedback loops and act on what you learn.”

While synthesising data to drive some evals is useful, I’m not yet convinced about synthetic user data: humans are diverse and weird. But it’s inevitable that automated feedback loops and experimentation will drive massive change over the next few years. So not enriching product sense with data will become even more untenable as Plan, Do, Check, Act cycles accelerate.

The nature of the feedback might change. The method of collection will too. Those who successfully bake learnings into their product process will thrive.

Be Flexible

“Sometimes, even great plans don’t work. Commit to action, but don’t get too attached to specific ideas. Be firm on your destination and flexible on how to reach it.”

In a world of expensive, slow development cycles, pivoting was hard in most environments. AI makes flexibility easier to practise on a daily basis. In my current project alone, we’ve iterated rapidly through three architectures to find the right balance between agents, tools and skills, and applied a ShadCN-powered design system retrospectively. Yet, because tweaking features costs less than it used to, holding a vision in focus becomes even more important.

In a world full of builders, what will stand out is the ability to solve real problems flexibly and elegantly.

Make it Happen

“Be proactive. It takes skill to apply product thinking in real-world environments. Push through obstacles to deliver value to your customers.”

The principle hasn’t changed. Can you read a situation, decide what’s worth doing, and get it done fast and well? While AI is reducing the cost of build and development, it’s unlikely to make us more decisive, politically astute, or comfortable turning ambiguity into action.

Three years ago, we battled procurement timelines, bureaucracy and brittle release processes. AI-flavoured obstacles (model behaviour, evaluation, AI-shaped governance) will begin to take their place. None of it alters the underlying point: being an effective operator has never been more important.

Bringing it all back home

As the cost of building falls, making ‘stuff’ will be commoditised. What differentiates the best from the rest will be insights, operating chops and a relentless focus on outcomes over outputs.

Clarity on what’s worth building, thoughtful design where it counts, feedback-powered iteration, blending talents and insights, effective execution: these will continue to be critical skills for a long time to come. So, while the product world has moved significantly since I wrote these principles, I’m pleased (and relieved) to say that they still feel incredibly powerful. I’ll certainly keep reading them every day.

---

Finding Product Breaks valuable? Please consider sharing it with your friends - or subscribe if you haven’t already.