What Is Dilr Academy? The AI Tutor Explained

Ask a room of education technology buyers what an AI teacher is and you will hear ten different answers. For some it is a chat window with a syllabus bolted on. For others it is a marking robot, or a video library with a search box. The label has outrun the thing it describes, which is a problem when the behaviour it points to is now almost universal. The HEPI Student Generative AI Survey 2026, published by HEPI and sponsored by Kortext, found that 94 per cent of UK undergraduates now use generative AI to help with assessed work. Students have already decided that AI belongs in their studies. The open question, and the one a university programme director actually has to answer, is whether the AI they reach for teaches them anything.

That is the question this guide answers, and it answers it through one product. "AI teacher" is a category, and if you are shortlisting several of them against procurement criteria you want the AI teacher buyer's guide, which compares the field and sets out how to run a pilot. This piece is narrower. It explains what Dilr Academy is, how it works, and where it sits inside the wider DILR.AI stack, so that when the word "AI teacher" comes up in a meeting you can say precisely what one version of it does.

This guide is shipped by the team behind Dilr Academy, the AI tutor described here. For the consulting and delivery side of the same stack, see DATS, our senior-led AI methodology.

What is Dilr Academy?

Dilr Academy is an AI-native learning platform from DILR.AI: an AI tutor that builds interactive, multilingual courses on demand and teaches through them, rather than answering questions on request. Its defining behaviour is that it asks before it tells. Instead of returning a finished explanation, it questions the learner, renders live diagrams and simulations, and moves on only once the learner has shown understanding.

It is a product line inside the wider DILR.AI portfolio, not the company's lead offer for schools or universities, a distinction this guide returns to later. The difference from what Academy is most often mistaken for is easy to see. A general assistant will answer whatever you type and stop there; whether you learned anything is your problem. Dilr Academy is built the other way round. A lesson is a sequence of prompts, checks and worked ideas, closer to a one-to-one class than to a search result. The teaching is the point, and the answer is a by-product of it.

What does "asks before it tells" mean in practice?

It means the tutor withholds the finished answer and works towards it with you. When a learner meets a new idea, Dilr Academy poses a question, waits for a response, and adjusts what it does next. If the answer shows a gap, it narrows the idea and tries again; if the answer is sound, it advances. This is Socratic questioning applied consistently, the opposite of an assistant that hands over a polished paragraph the moment you ask.

The mechanism matters because a tool that simply hands over finished work invites the learner to submit it rather than learn from it, and to decide for themselves whether they have actually understood it. HEPI found that the share of students directly including AI-generated text in assessed work has climbed year on year, a pattern set by tools that answer rather than teach. A tutor that asks first is a different interaction by design: the learner responds before the machine helps, so the help builds on an attempt rather than replacing it.

What is inside a Dilr Academy course?

A Dilr Academy course is a real syllabus, not a single long chat. Each course is organised into modules, each module into topics, and each topic carries stated outcomes that say what the learner should be able to do by the end. Running through all of it is a one-to-one AI tutor, with Deep Dives for the points that need more time. The structure is what lets the platform track progress against something concrete.

The live catalogue is deliberately concrete rather than aspirational. The same on-demand course building sits behind the set that is live today: 12 interactive courses across six or more subjects, drawn from Dilr Academy and partner organisations. That is a curated library rather than an endless marketplace, which suits the platform's premise: the value is in how a course teaches, not in how many titles sit on a shelf.

What is generative UI, and why does it matter for learning?

Generative UI is the platform building the right interface for a concept in the moment, rather than showing every learner the same static page. In Dilr Academy that means validated, templated interactive widgets, not layouts invented on the fly. When an idea is best shown as a diagram, the tutor renders one; when it needs a chart or a tree, it draws that instead. The interface is assembled from components that have been checked, not a hallucinated screen.

For difficult ideas the platform goes further than text. Dilr Academy renders live diagrams and charts inline, using Mermaid, D3 and tree visualisations, and it uses animations that explain hard concepts step by step. A learner working through a systems idea can watch it assemble rather than read a description of it. The point of generative UI here is not novelty; it is that the explanation takes the shape the concept needs, and it does so from a validated palette rather than by improvising a layout the platform has never rendered before.

How does Dilr Academy know whether a learner has understood?

It tracks mastery rather than activity. Dilr Academy uses mastery tracking that adapts to how the individual learns, with the difficulty of the next step tuned by knowledge tracing, a technique that estimates how well a learner has grasped a skill from their pattern of responses. Because the tutor asks before it tells, it has a stream of genuine attempts to read, not just a log of pages viewed, so it can move on only once understanding shows.

This is the difference between a platform that reports engagement and one that reports learning. Time in the app and messages sent tell you a learner was present; a mastery estimate tuned by knowledge tracing tells you whether the material landed. For a programme director judging whether an AI tutor is worth adopting, that distinction is the whole game, the kind of call our AI placement diagnostic is built to make, and it is why the structured-syllabus design described above matters: you can only track mastery of an outcome if the course states the outcome in the first place.

Where does Dilr Academy sit in the DILR.AI stack?

Dilr Academy is one line among several at DILR.AI, and it is honest about not being the lead offer for every buyer. The company's main enterprise revenue lines are its enterprise voice AI platform and its senior-led AI consulting and delivery work. Academy is the learning line, and for a school, university or corporate function the first conversation is usually about those rather than about Academy. Naming that plainly is more useful than pretending every product is the headline.

That is a deliberate position, not a hedge. DILR.AI's voice AI platform and its AI consulting and delivery work, delivered through the DATS methodology, lead for most buyers; this guide is one of a set of definitional pieces in our AI strategy writing that explain each line in its own terms. The same house that builds a private clinical extraction model that runs on a customer's own hardware and a governed agent work-management platform builds Academy: validated components, clear boundaries, and a preference for showing the working rather than asserting a result. If you are mapping where AI actually pays inside an organisation, our AI operating model work and an AI placement diagnostic start from the same premise, that a tool earns its place by moving a measurable outcome, and an AI execution office keeps it running once it is placed.

Who is Dilr Academy for?

It is relevant to anyone responsible for whether learning actually happens, from a university programme director to a corporate head of learning and development. The reader problem is consistent across those roles: learners have adopted AI faster than institutions have worked out how to guide them. A tutor that teaches, rather than one that answers, is one response to that gap, because it gives learners a sanctioned way to use AI that is designed around understanding.

The gap shows up in the survey data. The same HEPI survey found that only 36 per cent of students feel encouraged by their institution to use AI, and nearly two-thirds say assessment has changed significantly because of it. For a tutoring agency the enquiry and booking side of the operation is a separate problem, handled by our AI voice for tutoring agencies work, while Academy handles the teaching itself; if you are unsure which part of your own problem is which, the quickest route is to book a scoping call.

The sector's own verdict on AI tutoring is still forming, which is worth saying plainly to any buyer. The share of undergraduates submitting AI-generated text directly into assessed work has risen sharply across the survey's three annual editions, and the Education Endowment Foundation has opened a major research programme into how generative AI affects children and young people's learning and cognition, with first findings expected in 2027. Dilr Academy is a product you can adopt today, but the evidence base for AI tutoring as a whole is being built in parallel, and an honest guide names both.

As the HEPI report puts it: "Some 95% of students report using AI in at least one way and 94% say they use generative AI to help with assessed work." The behaviour is settled. What differs between tools is whether the AI teaches or merely answers, and Dilr Academy is built to teach.

Is Dilr Academy safe and multilingual?

Safety is built into how the interface is generated, not added afterwards. Dilr Academy renders in a sandboxed environment with tenant isolation, and because its generative UI is assembled from validated, templated widgets it does not present learners with hallucinated interfaces. The platform is also multilingual end to end, including right-to-left scripts, so a course can be delivered in a learner's own language rather than translated around the edges.

Adopting any AI tool in education carries duties on more than one side. The adopting institution remains responsible for how learner data is handled and for safeguarding, the kind of accountability the DfE, Ofsted and Ofqual expect of schools, inspectors and awarding organisations. Separately, the provider of an online service likely to be accessed by children carries its own obligations under the ICO's Children's Code. A responsible buyer treats an AI tutor the way it would any processor: it checks how data is handled, where rendering happens and who is accountable on each side. Dilr Academy's sandboxed, tenant-isolated design is meant to make that review straightforward rather than to remove any duty.

Is Dilr Academy a replacement for teachers?

No. Dilr Academy is a tutor and a structured classroom, not a substitute for a teacher or a lecturer. It handles the one-to-one practice, the questioning and the mastery tracking that are hard to deliver at scale, so a human educator's time goes to the things only a person can do. The honest description is an AI tutor that supports teaching, working through structured courses that state their own outcomes.

How many courses does Dilr Academy offer?

The live catalogue holds 12 interactive courses across six or more subjects, from Dilr Academy and partner organisations. It is a curated set rather than an open marketplace, which fits the platform's premise that the value is in how a course teaches rather than in the size of the shelf. Each course is a full syllabus of modules, topics and stated outcomes with a one-to-one AI tutor.

To go deeper, weigh Dilr Academy against the field in our best AI education tool guide, see how we place AI with an AI operating model and an AI placement diagnostic, or read our wider approach behind the stack.

Written by the Dilr.ai engineering team, practitioners who ship enterprise AI in production. Follow us on LinkedIn for shipping notes, or subscribe via the RSS feed.