AI training for teams doing serious work.
Hands-on AI training built around your team’s real work, whatever your sector: proposals, reports, analysis, operations and coordination across partners. Not generic demos.
How it works
Three phases. Teams can start with Foundations alone.
Before we startSet the board
Before the first workshop, your organisation shares what the training has to fit inside. Nothing is built on tools you can't use.
- Your technology stack and where your documents live
- Data and AI policies, and any constraints from funders or partners
- The list of approved and disapproved software, including AI tools
- Who is taking part and what they work on day to day
Phase 1Foundations
Three team workshops, about 2 hours each. Teams can start with Foundations alone.
How LLMs work
Why they predict rather than know. Context windows, memory, training cutoffs, and why the same prompt gives different answers.
Powers and failure modes
Where AI genuinely helps in your team's work, and where it fails. How to verify.
- !Drafting, summarising, restructuring and checking against requirements
- ??Invented citations
- ??Agreeing too readily
- ??Confident errors
- ??Drift across long documents
Setting up your environment
A hands-on session configuring the tools your organisation permits: projects, knowledge bases, instructions and file handling.
Phase 2Coached build
Over about three months, participants build their own workflows on live tasks with coaching. Examples: checking drafts against donor requirements, synthesising partner inputs, drafting reports. Available for a team or as individual coaching.
Phase 3Clinics #
Follow-up sessions a few weeks later to fix what isn't sticking.
What teams keep
- A shared library of workflows and prompts
- A verification guide
- Recordings of every session
Pricing
Foundations from $4,800 for a team. Coaching from $4,000.
Team programmes are quoted after a scoping call, based on how many people take part and how many workflows you want to build.
Discuss your team’s needs. Sessions are remote; timing is flexible.
