An AI tutor is a conversational agent that helps a student work through course material at their own pace. Ours answers only from the course it is attached to, cites the page, and is built to lead a student to the answer rather than hand it over.
All Education & EdTech →
Scalable one-to-one tutoring is the most-cited promise of AI in education, and products such as Khanmigo have shown the shape of it. The harder part is not explaining a concept; it is declining to do the work for the student. A general model asked to solve a maths problem solves it. A tutor has to notice that the request is the homework, and respond with the next step a student can take themselves.
That behaviour is a policy, and the policy belongs to the teacher. How much help on a graded assignment, whether a worked example is allowed, when a hint becomes an answer: those vary by subject, by age and by school, so they are configured per course rather than baked into a prompt nobody can see.
The second problem is grounding. An AI maths tutor or AI language tutor that invents a method the class was not taught confuses more than it helps. Retrieval over the course material, with a citation on each explanation, keeps it inside what was taught and lets a teacher check any answer in one click.
Teachers write the evaluation set: real questions from their students, plus the attempts to extract an answer that they have already seen. A tutor passes when its explanations match the taught method, its citations point to the right page and it holds the line on assessed work. We re-run that set whenever the model, the material or the policy changes.
Students are told in the first exchange that they are talking to an AI tutor and how to reach their teacher. For under-18 users the deployment follows the stricter defaults that GDPR, the UK Children's Code and, in the US, COPPA expect.
inside the LMS
assessed work? hint level?
this course only
next step, with citation
where the class is stuck
any welfare concern
An AI tutor is a conversational agent that explains course material, sets practice questions and responds to a student at their own pace. The difference between a tutor and a general chatbot is restraint: a tutor is grounded in a specific course, follows a policy on how much help to give, and is built to lead a student to the answer rather than supply it. The teacher sets that policy.
Some will try, which is why the policy check runs before anything else. On assessed work the tutor offers a hint or a parallel example instead of the solution, and requests for direct answers are visible to the teacher in aggregate. No design stops a determined student opening a different tool, so the aim is a tutor that is more useful than a shortcut, not one that polices.
Yes, with subject-specific checks. Maths answers are verified by computation rather than accepted from the model, and notation follows what the class was taught. Language tutoring adds speaking practice and correction against the level the course targets. In both cases the tutor works from the course material, so it teaches the method the teacher uses.
It runs inside the learning platform students already use, through LTI 1.3 for Moodle, Canvas, Blackboard and Brightspace, or as an add-on for Google Classroom and Teams. Sign-on, course membership and roles come from the LMS, so the tutor only sees the courses a student is enrolled on.
What we built, what stage it is at, and a link to the live thing wherever there is one.
yourOS is an operating system with an AI at its heart. You talk to it like a person, out loud or typed. It makes your calls, plans your days and works on your projects with a crew of agents, on your own private server. Our own product, built from the first commit; the waitlist is open.
Website, backend and app assets for Sonora, a free AI sound-therapy app on iOS and Android published by Cognitive AI.
Brand, website, prototype, apps on five platforms, the AI backend and the algorithm-intelligence pipeline for Podswap, a social growth engine that runs without a human in the loop.
In closed beta: Roxy, a private AI assistant with permanent memory running on infrastructure the user owns. Brand, site, product, backend and conversational design.
Live at coherence.com: brand identity generated with AI agents we directed, the motion system and website for the group behind Sonora, AboutAI, WakeUp and MudShadows. Agent work remains in build.
Live at domainbooth.com for DomainBooth, a premium domain brokerage: a new animated mark, every page redesigned and rebuilt with designed mobile layouts, and a custom CMS for its sales record, featured names, press and enquiries.
Start with one module, its material, and the questions your students already ask.
30-minute technical call · no deck