AI won't save a blurry vision

There’s a lot of pressure to use AI right now - and a lot of pressure to deliver faster.

These are probably the same pressure.

Without a clear vision, both will take you somewhere you didn’t want to go.

Stephanie at InsideProductOrg wrote something this week about why teams keep shipping products nobody buys. Worth a read. It got me thinking about something slightly different.

Teams aren’t skipping the hard thinking because they forgot it was important. They’re skipping it because they’re responding to pressure - and AI hands them something that looks like a response.

Sales are down. A competitor just shipped something. Someone senior doesn’t like a part of the product. Everyone else is “doing AI.” The instinct is to do something visible, fast. Throw a chatbot on the homepage. Spin up a prototype. Ship something - anything.

Reacting to negative signals isn’t new behaviour. AI just made the response faster.

Pressure starts at the top. A founder who’s frustrated at the pace of delivery reaches for AI tools directly, prototyping their own ideas rather than asking why the team is slow. The team then builds what the founder prototyped. Nobody stops to ask if it’s the right thing.

The problem with responding fast to blurry signals is that you compound the blur. Speed amplifies your existing direction - good or bad.

I’ve experienced this with my own projects. Work on one screen without thinking about how it fits with the others, and you end up with a product that feels different depending on where you are in it. Inconsistent UI, inconsistent logic, features that make sense in isolation but create friction as a whole. “Build me this” is easy. “Build me this so it fits coherently with everything else” is the hard part - and it’s the part that AI doesn’t automatically do for you.

The excitement around AI prototyping is real. Getting an idea into someone’s hands faster is genuinely useful. But ideas were never the bottleneck. PMs and users have always had loads of ideas. Sometimes just saying one out loud gets people excited. The hard part - will this work? does it fit? do enough people actually have this problem? - that’s still up to you.

Let’s talk about Lovable.

You’ll often see Lovable ($400M ARR through relentless shipping) cited as proof that just building fast works. But Lovable is a meta-product. They sell the tool that lets others build products. Their success doesn’t depend on any individual product built on Lovable succeeding. It depends on the platform being useful - and it clearly is.

WordPress made publishing faster. It didn’t make your writing better. Substack made newsletters easier. It didn’t make your ideas more interesting. Lovable lets you build products faster. It won’t tell you if your idea is good.

That’s not a criticism of Lovable - it’s just what platform tools do. You still need to decide WHAT to build - and determine if it’s a viable product.

So what’s the actual anchor?

A clear product vision. Not a broad statement that could apply to any company in your space.

Your vision needs to be specific. It should tell the right people that your product is for them - and help other people to understand why it’s not relevant to them.

An example I made up (it’s not a product I’ve created): “We help Python engineers write unit tests in a tenth of the time”

This sentence tells you who the product is for - what it does - and the measure of success. It’s also specific, and therefore excludes people - which is the part most teams are afraid to do.

Vision isn’t usually a “one and done” type of thing. It can take a few tries. You might pivot away from a failing vision - and that’s often the right call. But the companies that keep chasing every signal, every trend, every “we need to do AI” moment - they don’t fail because they moved fast. They fail because there was nothing to anchor against.

AI is a pressure multiplier. If the pressure is pointing in the right direction, it does good work. If it isn’t, it accelerates the drift.

Your vision should tell future users how you’re going to make their world better. It only needs to be one thing. Just do that one thing better than anyone else.

Figure that out first. Then use AI.

Inspired by Why your team keeps shipping products nobody buys — Stephanie, InsideProductOrg.

Product Notes is free and lands every week. Subscribe to get it in your inbox.