From Feature Pileup to a Deliberate Path

Tatiana Jordan

I've been building more products with AI lately, and the shorter feedback loop has changed how I work.

I don't just mean generating screens or writing code faster. That's impressive, but it's probably the least interesting part.

I'm doing competitive analysis, looking for gaps in the market, creating personas, researching how different users actually work, mapping their journeys, and testing complete workflows against each of them. I'm comparing features and pricing across competitors, thinking through business models, defining what belongs in an MVP, identifying what could actually differentiate the product, and working through how I would launch it.

Then I'm building it.

And when something doesn't work, I can pivot and do it again.

Work that once moved across disciplines, meetings, and weeks can now happen in a more connected loop.

For most of my career, product development was full of handoffs. Research informed strategy. Strategy informed requirements. Requirements informed design. Design went to engineering. Somewhere along the way we'd test it, learn something, and feed that information back into the process.

Each discipline brought important expertise, but the process itself created distance between learning something and acting on it.

That distance is collapsing.

I can find something interesting in competitive research in the morning, question whether we're approaching the problem the right way, change the product, and test the new workflow against multiple personas that afternoon. I can discover that a competitor charges for something everyone else treats as table stakes, reconsider the pricing model, and see how that changes the launch strategy and feature priorities.

It's not just faster execution. The feedback loops themselves are getting shorter.

And the faster I work this way, the more valuable my experience feels.

AI can generate an incredible amount of work. What I'm bringing is twenty years of questions.

Who are all the people using this? How did they get here? What does their day actually look like? What other applications are open beside ours? What frustrates them today? What would their ideal experience be? What would give them an easy button?

What's the predominant action? What information matters most right now? What can we remove? What happens when there are 100,000 records instead of twelve? What happens when five different personas need to use the same workflow differently?

Why do they keep exporting to Excel no matter how many filters, columns, and controls we add?

And then there are the product and business questions. What is everyone else in the market doing? Where are they all making the same assumptions? What could we do that would make this product meaningfully different? If I wanted to build the thing that put us out of business, what would it look like?

Those questions aren't new. I've been asking versions of them my entire career.

What's new is how quickly I can follow where they lead.

That's also changing how I think about MVPs.

For years, one of the realities behind MVP planning was engineering capacity. We asked what we absolutely needed for launch, cut everything else, and created a phase two list that everyone knew might never happen.

That constraint created useful discipline. I don't want to lose that.

But I'm starting to question whether “minimum” should carry quite as much weight when the economics of building have changed this dramatically.

If I can add something genuinely valuable in hours instead of weeks, why wouldn't I?

That doesn't mean adding features because they're cheap. If anything, the ability to build almost anything makes judgment more important. AI will happily help us create a bloated product incredibly efficiently.

But now I can cut something because it doesn't create enough value, rather than because we ran out of engineering capacity. And I can add something because it changes the experience enough to matter.

Maybe onboarding can actually be great at launch instead of something we fix later. Maybe we can support the workflow users would otherwise leave our product to complete. Maybe we can include the capability that makes someone wonder how they ever worked without it.

Maybe MVP becomes less about the minimum we can build and more about the minimum experience worth switching for.

That's a much more interesting constraint.

And it changes where I want to spend my time.

If AI gives me back weeks that would have gone into production, I don't necessarily want to use those weeks to produce even more things. I want to use some of that capacity to understand the problem better. Talk to more people. Test more scenarios. Explore the weird idea. Look harder at the market. Question the business model. Try three approaches instead of committing to the first plausible one.

In other words, I want to spend the productivity gain on judgment.

I don't think this eliminates the need for product managers, designers, researchers, engineers, or any of the other people involved in building serious products. The complexity of what we build hasn't disappeared, and neither has the value of different perspectives.

But I do think the boundaries between those activities are becoming much more fluid.

That breadth makes the shift especially useful to explore.

AI is giving me the ability to move from understanding a problem to shaping the strategy to building the experience to testing it in a loop that's tighter than anything I've experienced before.

The limiting question used to be:

What can we afford to build?

Increasingly, I think it's becoming:

Of everything we can now build, what's worth building?

That's where experience can help: deciding what deserves attention.

Back to Perspectives