Skip to content

How to check what Ulern teaches

The walkthrough, 1 minute 22 seconds, captioned. Real Ulern screens from a demo academy we keep for this; the people in it are invented.

A school that puts its name on what an AI tutor teaches is entitled to two answers. Where does what it teaches come from? And how do you know what each participant has actually learned? This post walks through how Ulern answers both, using the video above.

Everything on screen is the product, recorded in October 2026 in a demo academy we keep for this. The people are invented, and so are the participants’ answers (given by simulated participants). The knowledge base, its sources and every lesson in it are Ulern’s own.

A knowledge base the supervisor can open

Ulern teaches from a knowledge base, and the supervisor can open it. Browse shows the subjects, their modules and their topics. Each topic holds its notes, and each note says where it came from: a document, a web page, general knowledge, or a person who wrote it. Follow a note to its origin and you get the page it was read from, and when.

In the demo academy, the knowledge base was built from public pages, one subject with 52 topics. A school can also build it from its own documents, and then a note traces to the document, the page and the passage it came from. History keeps the record: every addition, what it added, and when.

The AI tutor teaches from it

The AI tutor (on screen it is called Uli) teaches from the notes on each topic, on the participant’s own situation. In the video, an operations manager opens with a routing decision from his work, and every question that follows is about that decision. Each answer is checked before the session moves on.

The same trail runs from the other end. A learning target in a plan’s milestones opens to its origin: the topic, its notes, and the pages they were read from. So a supervisor can check what a milestone will teach before anyone is invited.

A school that wants the AI tutor to keep strictly to its sources can choose “Only the sources”. Then the AI tutor teaches only what the notes say. A topic with no notes is skipped and filed as a gap, rather than taught from what a model happens to know. And the references, dates, amounts and quotations in a turn are checked against the notes before the participant sees them.

What the supervisor sees of each participant

For each participant, the supervisor sees where they stand on every plan. Where the school opens participants’ work to their supervisors, the supervisor can also read a session itself, read-only, with each answer beside the right one.

Mastery comes with its evidence. A participant’s report lists every topic with the level it aims at, the participant’s band (familiar, competent or mastered), the confidence behind it, and the number of answers it rests on. There is no single score to take on trust. Across a plan, the report shows mastery by module and who is stuck, so the supervisor’s time goes to the participants who need it.

Why it works this way

An AI that teaches can be wrong, and saying it is careful does not make it so. What makes it checkable is a record: the knowledge it stood on, where that came from, and the evidence of what each participant did with it. Ulern keeps that record and puts it in front of the people responsible for the teaching.

For the making side of the same demo academy, see how a capstone works in Ulern. If you want to check it against your own material, book a demo.

← All posts