The people who need our product most are convinced they don't need it
building aisa.to (AI skills assessment) and we keep hitting the same wall.
the people who score lowest on verification and critical thinking are usually the ones who tell us 'I already know how to use AI.' meanwhile, someone who says 'I'm probably not great at this' walks in and demonstrates really sophisticated evaluation habits.
it's Dunning-Kruger applied to AI skills — the gap between self-reported skill and observed skill is massive. and it creates a real go-to-market problem: how do you sell a mirror to someone who already thinks they look great?
what's working (slowly): positioning it as 'see how you compare' rather than 'find out if you're good.' competitive framing gets curiosity through the door. once they see the actual breakdown across 11 criteria, the conversation changes completely.
anyone else building products where your ideal customer is fundamentally resistant to the thing you're offering? curious how others navigate that.
Yes I have experiencing this too. I was inspired to help build an app to mentor younger people leaving university or early in their career. I heard so many saying that they didn't know what to do or how to find out. Yet once built and I recruited about 50 beta testers the very target market who need it never finished the assessment as it took too long. They wanted a 5 minute solution. I ended up pivoting the offer. Ironic
I think the difficult part is that people rarely look for solutions to problems they do not believe they have.
The challenge is not necessarily convincing them that AI matters. It is helping them see the gap between perceived skill and observed behavior without making them feel judged.
In that sense, awareness becomes part of the product itself. Once someone can see the gap clearly, the conversation about improvement becomes much easier.
Invisible problems are often harder to solve because they first have to become visible.
This is such a sharp observation — and honestly one of the hardest GTM problems to solve.
You’re not really selling a tool, you’re challenging someone’s self-perception. That’s always going to trigger resistance.
What you’re seeing lines up perfectly with Dunning-Kruger effect — the people who most need the feedback are the least likely to seek it out voluntarily.
The “see how you compare” angle is smart because it shifts from judgment → curiosity + competition. A couple of other angles I’ve seen work in similar situations:
Make it outcome-driven, not skill-driven
Instead of “are you good at AI?”, frame it as
→ “are you getting the best outputs from AI?”
People care more about results than self-awareness.
Use soft entry points
Quick, low-stakes hooks like:
→ “Most people fail this 2-minute prompt test”
→ “Can you spot the hallucination?”
Let them experience the gap before labeling it.
Third-party validation > self-reflection
People resist internal critique but trust external signals:
→ benchmarks
→ percentile rankings
→ “top 10% of users do X differently”
Sell to teams, not individuals (initially)
Managers are much more open to “assessment + improvement” than individuals who feel judged. Then it trickles down.
Turn it into a status signal
If passing your assessment becomes something people want to show off, the dynamic flips completely.
You’re basically trying to sell awareness before improvement, which is always uphill. The trick is sneaking awareness in through curiosity, competition, or status — not positioning it as a correction.
Also, if you ever need help stress-testing real-world usage (especially around how people actually use AI vs how they think they do), I’d be interested in collaborating or contributing on the debugging/implementation side.
WhatsApp: +1 (361) 332-6512