Education
A tutor that adapts to the learner, not the other way round.
An AI learning assistant that combines language models with learner data and structured teaching logic to diagnose gaps and personalise support.
- Project
- Adaptive AI learning assistant
- Client
- Education & training programme
- Capabilities
- Generative AI · Personalisation · Evaluation · Product engineering
- improvement in measured learning competency
- +23%
See it work
adaptive-tutor
interactive demo · scripted example
Ask as a student
- 01
Learner profile
—
- 02
Detected gap
—
- 03
Next action
—
- 04
Response
—
The problem
Generic AI chatbots answer questions. They don’t teach. They can’t tell whether a learner is missing a prerequisite, repeating the same mistake, or just needs a different explanation, and they can confidently explain things wrongly.
Our approach
We combined a language model with the parts a good tutor actually relies on:
- A learner model that tracks what each person has mastered and where they repeatedly struggle.
- Structured educational logic that decides what to do next, such as explaining, giving an example, asking a check question or revisiting a prerequisite, rather than leaving every decision to the LLM.
- Outcome measurement focused on learning competency, not just user satisfaction.
The result
Learners using the assistant improved measured competency by 23%: support that adapts to where each person is stuck, rather than a chatbot that answers whatever it is asked.
Client details are anonymised. Demos are illustrative, built on synthetic data and fictional brands.