AI Training
Help your team understand AI and implement it in the work they already do, deliberately and with a shared standard for what good looks like.
- Who it is for
- Marketers, analysts, ops and operations leads. No engineering background assumed.
- Format
- Live cohort, online or in person. Hands-on throughout, built on your own material.
- Duration
- Six sessions across six weeks, or a compressed three-day intensive.
Most teams are not short of enthusiasm
By the time we arrive, most people have tried AI. Some are getting real value. Others quietly gave up after a few disappointing answers and concluded it was overhyped. The gap between those two groups is almost never intelligence or seniority. It is method.
The people getting value have learned, usually by accident, how to give a model enough context, how to break a task into steps, and how to check the output. That is teachable in weeks rather than months, and it is what this programme teaches.
Built on your work, not a case study
Generic training produces generic results. People nod along to an example about a fictional retailer and then cannot map it to Monday morning.
So we build the exercises from your actual material: your briefs, your reports, your tickets, your tone of voice. By the final session each participant has automated or accelerated something they genuinely do, and can show it to the rest of the team.
Judgement, not just prompting
Prompting is the easy half. The harder skill is knowing when not to use a model at all, how to tell a confident wrong answer from a right one, and what needs a human signature before it leaves the building.
We spend real time on failure modes, on where models fabricate, and on how to build a check into a workflow so mistakes surface early and cheaply.
Indicative outline
Nothing here is delivered off the shelf. Every programme is rebuilt around your tools, your data and the work your team actually does, so the exercises use your material rather than a generic case study.
- 01Foundations and mental modelHow these systems actually work, what they are good and bad at, and where the risk sits.
- 02Context and prompting in practiceStructuring a task properly. Participants rewrite a real piece of their own work.
- 03Checking the outputFailure modes, hallucination, and building verification into a process.
- 04Retrieval on your documentsStanding up a small assistant grounded in the team's own material.
- 05Automating a real workflowEach participant picks a recurring task and builds it end to end.
- 06Capstone and handoverPresent what was built, agree team standards, and set the internal policy.