AI & LLMs
Teaching people to use these tools on the work they actually do.
Practical AI training built from your own documents and recurring tasks.
Delivery. Scoped to defined teams and tools, with a follow-up check.
Leaders who have bought licences, and the teams expected to do something with them.
Licences are barely touched, people are inventing their own habits, or an assistant is producing confident answers nobody checks.
What we teach
The work, not the product
The report written every Monday and the email answered forty times a week. A finance team learns more from ten minutes on their own reconciliation than an hour of feature tour.
Prompting that holds up
Context, examples, structure and iteration — taught as a habit rather than a list of tricks that stop working next model.
Verification
How to check an answer, what these tools get confidently wrong, and which tasks must never be taken on trust.
Where not to use them
The judgment that separates a useful team from an embarrassed one: what stays with a person, and why.
Tools you already own
Microsoft Copilot, Anthropic and OpenAI, taught on the platform your business has actually licensed.
Data handling
What may go into which tool, in plain language, tied to the rules your business has set.
Role-based sessions
Finance, operations, HR, sales, engineering — different work, different prompts, different risks.
Follow-up
A check weeks later on whether the habit formed, which roles it never reached, and what needs another pass.
On your work, or it does not stick.
- 01The real task
The report written every Monday, the email answered forty times a week, the document nobody wants to start.
- 02A usable result
In the room, on their own material — which is the moment the training lands.
- 03Verification
How to check it, what these tools get confidently wrong, and what must never be taken on trust.
- 04The habit
A check weeks later on whether it held, and which roles it never reached.
Judgment calls, anything with legal or employment consequence, and anything where being confidently wrong is expensive. Knowing where not to use these tools is most of what separates a useful team from an embarrassed one.
We run role-based sessions on the tools you already own — Microsoft Copilot, Anthropic and OpenAI — built from your own documents and recurring work, with verification taught as a habit and a follow-up weeks later on whether it held.
Built on the work in the room
Bring the actual documents, the actual queue, the actual report. The moment someone gets a usable draft of a thing they were dreading, the training has landed — and they will keep going without being asked.
Verification is the skill that matters
Anyone can get an answer. The teams that get value are the ones who know which answers need checking, how to check them quickly, and which tasks they should not hand over at all.
Attendance is not adoption
A check a few weeks later tells you whether the habit formed, which roles it never took in, and whether the tool was the right choice at all. Sometimes it was not, and that is worth knowing before the renewal.
What you keep
The code and the data·Your existing relationships·Approval and control·The ability to stop·The off switchWhat that means
Start with a 45-minute briefing.
No pitch. We’ll map your situation against what actually works and tell you honestly where to start.