Train each team for the work they actually do.

We train leaders, managers, commercial teams, engineers, and risk-sensitive functions using their own tasks, documents, decisions, and quality standards.

Plan a programme

A role-specific curriculum with a shared safety baseline.

Pick a role and read its course. A shared baseline runs through every column; the rest of the modules exist because a leader deciding where AI belongs and an engineer reviewing generated code are not learning the same thing.

Module · role
How these systems behave
Instructions and context
Verification and judgment
Rules of use
Direction and portfolio
Workflow redesign
Agent and copilot operations
Research, content, and CRM
Context engineering and MCP
Confidentiality and controls

Leaders

7 of 10 modules · 4 shared by every role

Choose where AI should change the organisation, what value to expect, and what must stay under human authority.

  1. How these systems behaveWhat the models are actually doing, where they fail, and why confident output is not the same as correct output.
  2. Instructions and contextWriting instructions that work in normal use by including context, constraints, examples, and reusable patterns.
  3. Verification and judgmentChecking output against sources, knowing when to distrust it, and keeping the decision with a person.
  4. Rules of useWhat may enter a prompt and what may not: data classes, privacy, copyright, and disclosure.
  5. Direction and portfolioChoosing where AI should change the organisation, sequencing investment, and metrics that separate adoption from value.
  6. Workflow redesignDecomposing a team process, deciding which steps move to AI, and setting the review points that keep quality owned.
  7. Confidentiality and controlsBias, records, human oversight, and escalation when output affects people or money.

Still theirs afterwardsAn AI direction brief and a prioritised portfolio

Training is useful only when it changes daily work. Every programme leaves teams with approved workflows, review methods, and materials they can keep using.

Practice on real work, not generic exercises.

Participants bring recurring tasks such as reports, research, and campaign briefs. We turn successful exercises into repeatable methods the team can use after the session.

Brought to the lab

The weekly report

Three hours every Friday across five tools, assembled by hand, with no consistent review.

  1. Collect from the systems of record
  2. Summarise against last week
  3. Verify every number at its source
  4. A person signs the conclusion

Left withA shorter, verified reporting workflow with a human decision at the end

Brought to the lab

Client research

Open-ended browsing, scattered notes, and a brief that repeats whatever the last search returned.

  1. Frame the question before the tool
  2. Demand sources for every claim
  3. Compare against what the team already knows
  4. Synthesise with citations attached

Left withA cited research brief the account team can trust

Brought to the lab

The campaign brief

A blank page, inconsistent inputs, and a tone that drifts with whoever wrote it last.

  1. Start from the real positioning
  2. Generate against the audience, not the product
  3. Review claims, tone, and originality
  4. Approve one direction, then vary it

Left withA reusable briefing method that protects the brand

How a programme runs

We map the work, teach a shared baseline, run role-based labs, document the methods that work, and support adoption after the sessions end.

  1. Map the work

    We interview the people doing the work, identify recurring tasks, understand current tools and constraints, and choose where training can create useful change.

    RemainsA shortlist of tasks with clear value

  2. Set the common baseline

    Everyone learns how the systems behave, how to instruct them, how to verify them, and how the organisation expects them to be used.

    RemainsA shared vocabulary

  3. Run role-based labs

    Each function practises on its own documents, decisions, workflows, and quality standards — not classroom examples.

    RemainsTested workflows

  4. Build working methods

    The exercises that worked become reusable prompts, workflow cards, templates, and review gates.

    RemainsPlaybooks in the team’s own files

  5. Support adoption

    Champions, office hours, and usage checks help teams keep using the approved methods during normal work.

    RemainsA 90-day adoption review and usage evidence

Which teams need to use AI more effectively?

Tell us the functions involved, the tools they use, and the tasks they need to improve. We will propose the right modules, practical labs, reusable materials, and adoption checks.

Plan a programme Or talk directly