Most career/personality tools follow the same pattern: fixed questions → fixed scoring → fixed report. It's efficient to build, but it treats every user like they fit into 16 boxes.
We wanted something closer to how a good mentor actually works — asking follow-up questions based on what you just said, probing deeper on ambiguous answers, adjusting the whole assessment path in real time.
So we're building PotenAI on adaptive AI conversations instead of static forms. The technical challenge has been fascinating: keeping the conversation focused, avoiding generic LLM output, and turning open-ended answers into structured, comparable insights.
We're still early and would love thoughts from other builders — especially anyone who's worked on structured extraction from open-ended conversation, or AI products in the assessment/EdTech space.
Following along or want early access? → nopick.site
Top comments (0)