Product Gap or Positioning Gap? Let AI Tell You Which One You Have

A customer asks ChatGPT which project management tool handles recurring tasks well. The model names two competitors. Your product has handled recurring tasks for two years. It just never comes up.

Inside the company, this looks like a missed feature. Someone opens a ticket. Eventually it reaches a roadmap review, gets prioritized against a dozen other requests, and maybe gets built as though it never existed.

That is the expensive version of a mistake we built Egaki's product solution to catch before it reaches engineering.

Two problems that look identical from the outside

A missing recommendation can mean one of two very different things.

The first is a real product gap: customers want something you genuinely do not offer, and competitors are being recommended because they built it and you did not.

The second is a positioning gap: you already built it, but AI does not know, does not trust the claim, or associates the capability with a competitor instead. The product is fine. The story around it is not reaching the model.

Treated the same way, both become a feature request. Only one of them should.

Why this signal arrives earlier than anything else you track

Support tickets and sales call notes capture what customers say after they have already evaluated you, often after they have already chosen a competitor. AI discovery captures the moment before that: the question itself, asked before the customer has committed to anyone.

OpenAI's shopping research shows how far this has already moved: the system asks clarifying questions about preferences, constraints, and budget before it ever generates a recommendation. The model is actively building a shortlist based on stated needs, well before a customer lands on anyone's website.¹ If your product is capable of meeting that need and the model does not know it, you are excluded from a shortlist you should have been on, at the exact moment it gets built.

A simple way to tell the two apart

Before a missing recommendation becomes a product requirement, it is worth checking which of four situations you are actually in:

• AI already recommends you correctly. Nothing to do here except keep monitoring, since models and competitors both shift.
• AI never mentions the capability, and customers never ask for it either. That is a low-priority signal, not a gap worth resourcing yet.
• AI undersells or omits a capability you already shipped. That is a positioning problem. The fix is content, documentation, structured data, or clearer claims that a model can pick up and trust, not new code.
• AI consistently surfaces a need your product does not meet, across multiple personas and contexts. That is the one worth taking seriously as a real product gap.

The expensive mistake is treating the third case like the fourth. Teams that skip the diagnosis end up building something that already exists, because nobody checked why the model did not already know about it.

We built Egaki to catch this before the visit even happens: what the customer asked AI, what the model told them, and only then what they did once they reached your site. That ordering is the whole point. Running this diagnosis before a request reaches a roadmap turns a quarter-long debate into a five-minute lookup.

Positioning gaps are often the larger share

In practice, the second category (a real capability the model simply is not associating with your product) shows up more often than teams expect. A feature can be fully shipped, documented in a changelog, mentioned once in a blog post, and still be effectively invisible to a model that has never seen it reinforced across enough sources and contexts to trust the claim.

The fix for that case sits with marketing and content, not engineering: clearer product pages, comparison content that states the capability plainly, and documentation that uses the same language a customer's prompt would use. Our marketing team solution and the product side described here answer two ends of the same question: what does AI say about us, and is that gap in the product or in the story.

Prioritizing by who is actually asking

Not every missing recommendation carries the same weight. A capability a small segment asks about occasionally is a different priority than one an entire high-value persona keeps running into across every decision stage.

This is where persona and context matter as much as the request itself. The same missing feature can be a rounding error for one audience and a deal-breaker for another. Knowing which persona is asking, at which stage of their decision, before deciding whether to build, is the difference between a roadmap driven by evidence and one driven by whoever complained most recently. We built our persona-level breakdowns to surface exactly that weighting, so a request from your highest-value segment does not get the same priority queue as a one-off from someone who was never going to buy.

Put the diagnosis before the decision

Engineering time is too expensive to spend resolving a story problem with a product solution. Before a missing AI recommendation turns into a sprint, it is worth answering a narrower question first: is the gap in what we built, or in what the model knows about what we built.

At Egaki, we connect what customers ask AI, what the model tells them, and what they do once they reach your site, giving product and content teams the same evidence instead of two competing guesses. Read the full framework in the Egaki Playbook.