The hardest user research call we had was with someone who scored in the top 10%

they didn't believe us.

we showed them the evidence, walked through each criterion. they kept saying "but i'm just doing what makes sense." that's the whole problem with AI skills — the people who are actually good at it can't tell you what they're doing differently because it's become intuitive.

that single conversation changed how we structure the output. we stopped just giving scores and started showing the specific behaviors that earned them. turns out people don't just want to know where they stand — they want proof of what they're actually doing that works.

still early but this one change made the post-assessment conversations way more productive. people stopped arguing about the score and started asking "wait, what exactly did I do when I..."

building at aisa.to