Book a demo

Two AI platforms, and the engine that writes the courses.

Talk & Learn teaches a chapter by talking with one student. The Case Study Platform runs the Socratic case method with a room full of them. Both are authored in the same place, by your own curriculum team, in hours.

Book a live demo See a course get built
2 platforms
One conversational mentor, one Socratic case method
Hours
To author a full course, not weeks
Every session
Ends in a rubric-scored report
{{ heroMentor }}
{{ heroChapter }}
Live
{{ m.text }}
{{ heroRubric }}
This page answers six questions, in order
01 What are the two platforms? 02 Who writes the courses? 03 What happens in a session? 04 How does the case method work? 05 Is it really teaching, and is it safe? 06 What would it take to run it here?

What are the two platforms?

One teaches a single student through conversation. One puts a whole class inside a business case and argues with them. Pick either to see it run.

{{ platName }}
{{ platClaim }}

{{ platBody }}

{{ pt.t }}
{{ platMentor }}
{{ platContext }}
{{ platBadge }}
{{ m.text }}
{{ platRubric }}
{{ platFoot }}

Who writes the courses?

Your curriculum team does, in an afternoon. Chapters, objectives, rubrics and the mentor persona are authored in one place, and the same engine writes the case studies for the case platform.

Courses are built, not commissioned

This is the authoring console. Choose what you are making, then set the subject and who it is for. Everything on the right is generated from that.

{{ cbSubjectLabel }}
{{ cbAgeLabel }}
{{ cbNote }}
{{ cbTrackName }}
{{ cbTitle }}
{{ cbTime }}{{ cbTimeLabel }}
{{ c.n }} {{ c.t }} PromptWorkflowInputs
Rubrics mapped automatically {{ r.t }}

{{ cbFoot }}

{{ b.k }}
{{ b.t }}

{{ b.d2 }}

What happens in a session?

A written mentor teaches one chapter by talking with the student. The same conversation is the assessment, so nothing has to be tested separately.

Most digital learning splits the lesson, the practice and the test into three products. This keeps them in one dialogue. Four things make that work.

{{ p.t }}

{{ p.d2 }}

Four steps, about thirty minutes

The student is asked, answers, gets questioned, then revises. Every session ends with a report card. Step through it.

This is the shape of every session, whatever the subject. Step through it.

{{ stepTitle }}

{{ stepBody }}

{{ stepOut }}

Or watch two real sessions instead

Played back exactly as a student saw them. Press play on either.

And the mentor is written, not generic

Character is set at course design time, along with tone and the kind of example used. Here is the same weak student answer, handled by three mentors.

Student says: “I think it just depends.” The mentor replies

{{ prLine }}

{{ t.t }}

How does the case method work?

Three phases. The student forms a view alone, defends it against a Socratic AI, then sees the reasoning scored. The AI asks questions. It does not hand over answers.

The facilitator is the scarce part of case teaching. This is the part the platform replaces, so a case class no longer needs an expert in the room.

{{ phaseTitle }}

{{ phaseBody }}

{{ phaseOut }}
Case brief · post-graduate
The Automated Shortlist

A hiring team ships a screening model. It rejects a candidate the head of engineering wanted to interview. You decide what happens next, and you write down why before you talk to anybody.

Your initial position, logged

“Keep the model. The track record data predicts performance and the manager is going on instinct.”
{{ m.text }}
Asks, never answers · the position moved on evidence
{{ b.name }}{{ b.band }}
Scored on the rationale, not on picking the answer the case author preferred.

Every case is scored on two layers

One layer never changes, so a student can be tracked across every case they ever do. The other is written per case and mapped by the platform when the case is authored.

Standard learning rubrics

Universal, tracked across every case.

{{ b.name }}{{ b.band }}
Case-specific rubrics

Written for this case, mapped at authoring time.

{{ r.t }}

Aligned to the lesson content, so a case can be marked against what you actually taught that term.

{{ g.k }}
{{ g.t }}

{{ g.d2 }}

Four students, one case, one moderator

Students share a single discussion. The AI moderates live, keeps it on the rationale, pulls in whoever has gone quiet, and still scores each student on their own. Play it.

Four students, one case, one moderator

Press play and watch the AI keep the group on task and pull the quiet student in.

AI moderator{{ gdMode }}
{{ p.n }}
Bars show share of the discussion. The moderator watches these in real time.
{{ gdClock }}
{{ l.who }}
{{ l.int }}{{ l.t }}
{{ m.t }}
{{ m.d }}

Every voice in the room still gets its own read

After a group session each student receives moderator feedback and a competency profile of their own.

Competency breakdown · Kabir S.

Bakery case, Monday losses

{{ c.n }}{{ c.v }}
AI moderator feedback

What the moderator saw

Strength
Brought the only costed number into the room and it changed the group's decision.
Watch
Spoke twice in ten minutes. High value, low volume. Needs prompting earlier.
Next
Open the next case. Give the first summary before others state positions.
Mastery
Unit cost and break-even: secure. Trade-off framing: developing.

Is it really teaching, and is it safe?

The objective is fixed and only the example changes. Every session is scored on reasoning, teachers and parents read the full transcript, and the guardrails act before any adult needs to.

First, the same idea reaches each student differently

Pick an interest and watch one objective land four ways.

Learning objective, fixed
Explain opportunity cost in your own words

{{ interestLine }}

{{ interestWho }}

Then every chapter leaves a report behind

Conversation quality and learning outcome are scored together. Move through the chapters and watch both the report and the long-term picture change.

Chapter
Chapter report · {{ chTitle }}

{{ chStudent }}

Strengths
{{ chStrength }}
Gaps
{{ chGap }}
Next steps
{{ chNext }}
Long-term skill growth

Across chapters 1 to {{ chN }}

{{ s.name }}{{ s.val }}
The pale marker is where this learner started in chapter 1.

Teachers and parents read the transcript, not a summary

Two dashboards, one record. Switch between them.

Two dashboards over the same conversation

Teachers see the class. Parents see their own child.

{{ s.t }}{{ s.d2 }}

And the guardrails, tested on the messages you are worried about

Pick a message a student might send. See exactly what the platform does with it.

{{ sfFlag }}

{{ sfReply }}

{{ c.t }}

What would it take to run it here?

You choose where it runs, including fully offline. Every model in the stack is replaceable, and the roadmap is short and specific.

It runs where your board needs it to run

The language model, the speech synthesis and the speech recognition are separate and replaceable. Tap swap to change one.

{{ l.k }} {{ l.v }}
Nothing above is hard-wired. If your state or board requires a specific model, it drops in.
{{ depTitle }}

{{ depBody }}

What lands next

{{ r.d }}

What changes for a school that runs this.

If you read nothing else on this page, read these four.

{{ w.n }}
{{ w.t }}
{{ w.d2 }}

See it run with your own subject.

A 30 minute walkthrough of one full chapter, start to report card.
A live group discussion with your own case study.
The teacher and parent dashboards, with real transcripts.
A written note on deployment, data handling and guardrails for your board.

Book a consultation

Tell us the grade and the subject. We will run the demo on that, not on a canned example.