T Shape: AI training built around each person's job

Role explorer

See what changes from role to role

Pick a role, a level and an industry. These are samples, but they show the idea: each practice task uses real work for that role in that industry, highlighted, and names the mistake the AI makes that the learner has to catch.

Role

Level

Industry

Operations leader in financial services, Foundations

Objectives

Spot repetitive work in your processes that AI can safely speed up.
Write prompts that produce consistent, checkable outputs.
Know which data must never leave approved tools under financial-services conduct rules.

A practice task

Have AI turn a week of account-change requests into a summary of what’s slowing the team down, then check its counts against the queue.

Real programmes are built from your own organisation's context, and a person reviews them before anyone starts learning.

One case, two vantage points

People who work together can learn on the same case, each at the depth their job needs. The team ends up with a shared way of talking about AI, which makes it easier to work on together.

Customer team memberUses AI to draft a reply to a complaint, checks it against policy and sends it.

Team managerReviews the same complaint with AI: tests the decision, spots the pattern across similar complaints and picks what to automate.

Why T Shape

Most AI training is written for nobody in particular

A typical course gives the chief executive and a new customer service hire the same videos, the same examples and the same quiz at the end. The examples usually come from someone else's industry. The quiz checks that people watched. It doesn't tell you whether anyone works differently afterwards.

Meanwhile, AI spending is racing ahead of training, and regulators have noticed.

The value of AI was never in the tools. It’s in the people using them, and that’s where we start.

US$2.7tn
Forecast worldwide AI spending in 2026, up 49.5% on 2025.
Gartner, Sep 2026
36%
of employees say they've had enough AI training. No better than in 2025.
BCG, AI at Work, Jun 2026
Aug 2026
EU regulators began supervising the AI Act's literacy duty: measures suited to people's knowledge, experience and context of use.
EU AI Act, Article 4, as amended

Consultancy depth, at software scale

Consultants can tailor a workshop, but they charge by the day and only reach the people in the room. We do the discovery that other platforms leave to you, and AI lets us do it for every role.

Question	Off-the-shelf e-learning	Consultant workshops	T Shape
Written for your organisation	No, one course for everyone	Yes	Yes
Who does the discovery	You, if anyone	The consultants	We do, with AI agents and a reviewer
Reaches every role	Yes, with the same content	Only the people in the room	Yes, with a programme for each role
Practice in your approved tools	Rarely	Sometimes	In every practice task
Keeps up as tools and rules change	When the vendor updates it	Until the workshop ends	Programmes are kept current

How it works

How a programme gets made

AI agents do the research, design and drafting. That's how every role gets its own programme without a training team writing each one by hand. People make the decisions along the way.

We talk to you first. An AI-guided conversation, plus any documents you share, builds a brief: what you want people to be able to do, which roles are involved and which AI tools you've approved.
You sign off the brief.
We look up the public context. Agents research the rules and trends in your sector, and what's publicly known about your organisation. Every source is dated and linked.
You can check every source.
We propose a detailed curriculum for each role and case study for each team.
Our reviewers check what the agents propose.
You read the draft and we finalise it. You can comment while it's a draft. Our reviewers make the final call on the curriculum.
A reviewer approves every lesson before release.
People learn. Each learner has a short private interview, gets a plan that skips what they already know, and works through it with Tee, our AI tutor.
Anything you make compulsory stays in.
They show what they can do. Practice happens with contextual realistic hands-on activities.
Learners choose what they bring back.

A sample curriculum

Every programme has the same shape: an introduction, then three parts. What goes inside each module is built around the learner. Pick a module to see how.

Orientation · Module 0

Introduction

What it coversHow the programme works and how it adapts to your role.

Built around the learnerBuilt around the learner’s role, organisation, team and the people they work with, and introduces the running case used throughout.

Understand · Module 1

AI Fundamentals

What it coversWhat AI is, the key terms, and what has changed recently.

Built around the learnerTerms explained through tools and conversations the learner already meets at work, including what colleagues mean when they say “AI”.

Understand · Module 2

Identifying Opportunities

What it coversWhere AI can add value in your work, and where it cannot.

Built around the learnerMapped against the learner’s own weekly tasks and decisions, showing which suit AI and which do not.

Understand · Module 3

How AI Produces Answers

What it coversWhy AI can be convincing and wrong, and when to trust it.

Built around the learnerExamples use questions from the learner’s domain, so they see which ones are safe to ask and which carry risk.

Apply · Module 4

Providing the Right Information

What it coversGiving AI accurate, current sources to work from.

Built around the learnerUses the learner’s own document types, systems and policies, including the version and source errors common in their area.

Apply · Module 5

Briefing AI Effectively

What it coversTurning a task into clear instructions that produce usable work.

Built around the learnerLearners write briefs for their recurring tasks and leave with a reusable library for their role.

Apply · Module 6

Evaluating and Improving Output

What it coversJudging quality against a standard and refining the result.

Built around the learnerOutput is scored against the quality standards their organisation or function already applies.

Lead · Module 7

Delegating to AI Agents

What it coversHow agents work, what they can do, and how to oversee them.

Built around the learnerAgent scenarios are drawn from the learner’s workflows, with oversight pitched at their level of accountability.

Lead · Module 8

Governance and Responsible Use

What it coversRisk, accountability, and the decisions that stay with people.

Built around the learnerApplied to their organisation’s policies, regulatory setting and approval lines, with grey-area cases from their function.

Lead · Module 9

Measuring Value and Planning Ahead

What it coversAssessing return, prioritising initiatives, and the final assessment.

Built around the learnerLearners size and pitch an initiative from their own work, framed for the audience they would present it to.

For learners

What learners get

Most of the learning happens in a conversation with Tee, the AI tutor. Tee explains things using examples from the learner's own job, sets practice tasks and gives feedback on the work that comes back.

Practice on a realistic company: Lessons are set in a made-up organisation that works like yours, with its own records, policies and people. Learners practise on realistic material, so there's no need to use real customer data.
A tutor that fits the learner: Learners choose how Tee looks and how it teaches. From then on Tee sits beside every screen, ready to explain or coach.
Skip what you already know: A short private interview and an optional skills assessment shape each person's plan.
Work in your own AI tools: Practice tasks come with instructions for the tools you've approved, so the skills carry straight over to real work.

Two levels

Foundations: For everyone. Using AI confidently and safely day to day: what it's good at, what to check, and what should never go into it.
Frontier: For people who want to go further once they've finished Foundations, where your organisation offers it. Longer workflows, judging the quality of AI output, and using it on the harder parts of the job. Still written for people who don't code.

For leaders

What leaders get

Sponsors and admins see how capability is building across the organisation, by role and team, without seeing anyone's scores. Admins get one console for adding people to cohorts and seeing who has finished. Small groups are combined with others, so a lone executive's results are never shown on their own.

Sample capability map, 240 learners Share at or above target

AI fundamentals 67%

Prompting and applied use 52%

Agentic AI 33%

Governance 58%

Workflow improvement 17%

Evaluation 12%

Design
Evaluate
Ideate
Contribute
Discuss
Recognise
Not yet
Your target

The percentage is the share of people at or above the target. Sample figures only.

Opportunities

From lessons to a list of real opportunities

While people learn, they notice where AI could help with their work, and what's stopping them from using it. T Shape keeps track of both, so a programme ends with more than completion numbers.

Learners note ideas and barriers as they go. The Idea Canvas sits beside every lesson. Learners work their ideas up with Tee, and nothing leaves the canvas until they choose to share it.
They share at the end of the programme. Ideas are shared with the learner's name, so good thinking gets credit. Barriers are always anonymous, so people can be honest about what's in the way.
Your leaders decide what happens next. The Opportunities and Challenges Builder groups everyone's ideas and barriers into themes. Your sponsor and admins tidy them up, mark each one green light, later or not now, and publish a map of value against readiness.

Value Readiness Triaging complaints by risk and urgency Checking supplier contracts against our terms Spotting at-risk renewals in account notes Pre-checking claims for missing evidence Assembling the monthly board pack

Sample only. The themes on your map come from your own people's ideas. Black circles are marked green light, yellow diamonds later. Higher means more value; further right means readier.

Team leaderboard

Teams are ranked by the ideas they've developed and shared, not by how quickly people click through lessons. Each learner sees their own score and roughly where they sit in their cohort. They never see anyone else's name or score, and a team only appears once five of its people have shared an idea.

Privacy

Private by design

Learning works better when people can get things wrong in private, so the permissions are built around that.

When a group is small enough that someone could be identified, such as a single chief executive, their results are combined with others or held back.

Information	Learner	Manager	Client admin	Sponsor
Conversations with Tee	Their own, stored for them alone	No	No	No
Individual assessment results	Their own	No	No	No
Progress and completion	Their own	No	With their name	By role and team
Ideas they share	Their own	With their name	With their name	With their name
Barriers they share	Their own	No	Anonymous	Anonymous
Capability across the organisation	No	No	By role and team	By role and team

FAQ

Questions people ask us

Which AI tools does it work with?

The ones your organisation has approved. Practice tasks come with instructions for those tools, and learners paste or upload what they made. There's no direct connection to the tools, so we only see what learners bring back.

Do we have to share confidential information?

No. We ask for the context that shapes the programme, such as your goals and AI policy, and send AI providers only what a task needs. Every practice task comes with made-up records, so learners don't need real customer data, and your documents are never used as public search queries.

How long does a programme take?

There's no fixed length. We estimate the effort for each role's objectives. If you have a time budget, we'll suggest priorities or a phased plan that fits it.

What if someone already knows this?

They can show it in a short skills assessment and skip ahead to harder material. Anything your organisation makes compulsory stays in their plan.

Who checks the content?

Automated checks run first. Then one of our reviewers approves every lesson before it's released. You can read and comment on the curriculum while it's a draft.

Is the certificate accredited?

No. It records what someone demonstrated in their programme. We don't claim external accreditation.

What do learners need?

A web browser and access to your approved AI tools. Lessons are text conversations with Tee, so there's nothing to install.

Our note

Why we've built this

Before T Shape, we spent years helping large organisations get value from AI. As consultants, we saw the same pattern on almost every engagement. The strategy got written. The tools got bought. Then the people side was handed to L&D, who either ran one course for everyone or brought in consultants for bespoke workshops.

The course was too generic. The person answering customer emails sat through the same slides as the finance director, and neither saw their own work in it. The workshops were tailored but costly, so they reached a few teams and stopped there. Either way, most people went back to their desks and kept working the way they always had.

That was never a problem with the people. It was a problem with the training. Knowing what AI can do in general doesn’t tell you what to do on Tuesday morning with the file in front of you.

So we left consulting to build the part we kept seeing go missing. T Shape teaches each person through their own role, their own tasks and the standards they already work to, with a tutor that knows their context. One programme runs from the board to the front line, so everyone gets training built around their job, not just the few teams a workshop can reach.

The tools don’t create the value. People do. We help every one of them get there.

If you invest in or partner with companies working on this, we'd like to hear from you. Use the form below and choose "Investing or partnering".

The T Shape team

Get in touch

Book a demo

Tell us about your organisation and we'll show you a programme built for a few of your roles. Use the same form if you're interested in a pilot or a partnership, or have security and data questions.

A person on our team reads every request and replies.

Or email us directly at hello@tshape.ai.

Your finance team and your call centre shouldn't take the same AI course.