Where AI belongs in your product
There’s an electric toothbrush on sale with “with A.I.” printed on the box. That’s roughly where we are. The label has become the feature, and it seems to sell more units.
The pressure on founders is real. If you’re not shipping something with an AI angle, it feels like you’re missing out, or worse, being left behind. So teams bolt an assistant into the corner of the screen, tick the box and move on.
I think that’s the wrong instinct, and I think it’s a product problem, not a technology one. The question isn’t whether to put AI in your product. It’s where.
The bar is high
To be clear, being sceptical isn’t the safe option either. When Klarna launched its AI assistant, within a month it was handling two-thirds of all customer service chats. That’s 2.3 million conversations, the work of 700 full-time agents, with resolution times down from 11 minutes to under two, a 25% drop in repeat enquiries, and an estimated $40 million profit improvement in a year.
That’s not a gimmick. That’s a whole business function rebuilt.
The bigger picture is that the large, “wide” platforms are now established, and that’s opening the door to a wave of “narrow” tools built on top of them. The infrastructure question is largely settled. The open question, and where the value now sits, is knowing which specific problem to point it at.
Assistants are the obvious answer, and that’s the problem
Design patterns are starting to settle, and the most common one is the AI assistant: a chat panel bolted onto an existing product. It ticks the box, and it can be useful. But most of them feel opportunistic. They’re a response to the moment, not to anything a user asked for.
My view is that as this matures, we’ll see much less “Powered by AI” and far more seamless experiences, where users don’t know or care that AI is involved. It’ll go from being an exciting new feature to something everyone expects. The products that win will be the ones where it’s doing something specific and invisible, not something general and loudly announced.
The balance to strike is between gimmick and value. That means starting with the experience, not the technology.
Two ways to find the right spot
Both of these are everyday product techniques. Neither is about AI, and that’s why they work. They find the problem first, and only then ask whether AI is the right tool to fix it.
Customer journey mapping
Map every interaction a customer has with you, from first contact to after they’ve bought, across every channel. For each stage (registration, onboarding, first session and so on) note down four things: what the customer does, which part of your service they touch, what they’re thinking, and how they feel.
The feelings row is the one people skip, and it’s the one that matters. It’s where you find that registration is a joy, onboarding is confusing and frustrating, and the first real session is good again. That dip in the middle is worth more than any amount of guessing about what to build.
Then, and only then, go through each opportunity you’ve found and ask whether AI makes it better. Some will be obvious candidates: shortening a training flow, surfacing the right template at the right moment, sorting a support queue. Many won’t be, and the discipline is in leaving those alone.
The speedboat technique
This is a group workshop with users or potential users. Draw a speedboat. Ask what’s pushing them forward and what’s holding them back.
The anchors are the things slowing customers down. The engines are what’s working and could work harder. Prioritise the biggest anchors and the most powerful engines, then build an action plan around them.
It’s a quick, cheap way to get candid input. And because it’s framed around a boat instead of your roadmap, people tell you things they wouldn’t otherwise say.
What this changes
Run either exercise and the conversation changes. You stop asking “where can we put AI?” and start asking “which of these specific problems, with evidence behind them, does it solve better than anything else?”
Usually the answer is one or two places, not ten. And they tend to be unglamorous: the pain point in onboarding, the repetitive triage nobody enjoys, the personalisation that was always too expensive to do by hand. They rarely make great launch copy. They’re usually the ones that show up in retention.
The founders I see getting this right have one habit in common. They treat AI as one option in a product decision, not the decision itself. They’re also, almost without exception, the ones with someone senior owning product. That’s because the way this goes wrong isn’t technical. It’s a roadmap driven by what’s fashionable instead of what the evidence says.
If this sounds like you
This is a lot of what I do. I work with founders as a fractional CTO and fractional CPO, giving them senior technical and product steering. In practice that often means running exactly these sessions, then making sure what comes out of them is buildable and worth building.
If you’re looking at an AI roadmap you’re not confident in, or you’ve shipped an assistant and can’t tell whether it’s earning its place, tell me what you’re working on. A short conversation is usually enough to tell whether there’s something real there.