Career guidance is rationed by counsellor hours. A student typically gets one short conversation, at the point in their life where the decision compounds hardest, and what follows is a college list ranked by somebody else's criteria rather than by anything the student said.
Edmertion puts the counselling conversation in front of the student as an AI they can talk to at any hour, and matches colleges against what that conversation reveals rather than against a published rank order. Built in 2023 as three connected pieces — the web experience, the mobile app and the portal behind both.
Counselling is a conversation, so the interface is one
A form asks a student to already know what matters to them. A conversation lets the system find that out — interests raised in passing, constraints admitted late, the subject they are actually good at rather than the one they are supposed to want. The interface choice is the product decision here, not a styling one.
Matching against a profile, not a rank
National rankings answer a question no individual student is asking. Matching runs against what the counselling session established — course, location, cost, entrance route, distance from home — so two students with the same score can correctly receive different lists.
Built in 2023, before the pattern settled
Conversational AI in a consumer product was not yet a solved shape when this was built. That put the engineering weight on the parts that do not go out of date: the student profile the conversation produces, the matching logic underneath it, and a portal that lets a human review what the system said.