How AI actually helps you make better product decisions

Most writing about AI in product management describes a future that doesn't exist yet. Predictive roadmaps. Autonomous prioritisation. AI that knows your users better than you do.

That's not where I am. Here's where I actually am.

The problem AI solved for me wasn't the one I expected

I didn't start using AI to make better decisions. I started using it because I had too many inputs and not enough time to make sense of them.

The moment it clicked was consolidating project files. Multiple docs, different formats, written at different times by different people — getting a coherent story out of that used to take most of a day. Getting the narrative right, finding the through-line, working out what to lead with. Claude cut that to an hour.

That's not a small thing. The time you spend wrestling with structure is time you're not spending on the actual decision. When that friction disappears, you think more clearly about what matters.

What changed about how I work

The practical shift was in presentations and narrative. Not the slides — the story. Working out what this product is really about, what the ask is, what the audience needs to believe by the end. That used to require multiple drafts, rounds of feedback, a lot of staring at a blank doc.

Now I bring Claude the raw material — the brief, the research, the scattered notes — and use it to find the shape of the argument first. The first draft is never right. But it's a real draft, not a blank page. That changes everything about how fast you can get to something worth showing.

The decisions get better because the thinking gets cleaner. When you can see your own argument laid out clearly, you spot the gaps faster. You know what you don't know. You go into the room with a sharper point of view.

Where it doesn't help

AI is only as good as what you put in. If your inputs are vague — half-formed briefs, assumptions dressed up as facts, a brief that hasn't been properly challenged — the output is vague too. Faster, but vague.

The judgment calls still belong to you. Which problem is worth solving. Whether the user insight is real or assumed. Whether the business case holds up. AI won't tell you any of that. What it will do is help you structure your thinking quickly enough that you have more time to make those calls properly.

That's the honest version of AI in product work. Not a replacement for thinking. A way to get to the thinking faster.

The practical takeaway

If you're not using AI in your product workflow yet, don't start with the big stuff — roadmap prioritisation, market forecasting, the things vendors promise it can do. Start with the messy middle. The synthesis. The narrative. The moment where you have all the inputs and need to make sense of them fast.

That's where the time goes. That's where AI actually earns its place.

I've written more about the specific process changes in what changes when you plug AI into your product process, and a broader take on the strategy side in AI won't build your product.