Small-group coaching
2 to 3 colleagues work on a real task of theirs and leave with it done.
75 minutes
Workshops and team programmes
For teams whose views on AI range from refusal to daily use. We begin with what the team values, practise on its real work, and end with decisions the team can stand behind.

Teams of 5 to 15 in regenerative agriculture, restoration, conservation, policy and other impact work.
Team leads who want the whole team to arrive at a shared position on AI.
Organisations writing AI guidelines that need them to work in everyday practice.
Two afternoons together. A blind read of two versions of a report in the team's own genre, one written by people and one by AI. Possible futures for the team's work. What the team values and will not hand over. Questions answered from what the room actually asks. One small experiment per person.
Online sessions every few weeks on real work: one task done by hand and with AI, then compared; a literature search checked against a hand search for what gets missed; summaries checked for what they flatten, especially from local and non-Western sources.
A closing session where the team decides what to build, pilot or leave alone, and writes down why.
Optional: advice on organisational AI guidelines, a session for all staff, and ongoing coaching for colleagues.
2 to 3 colleagues work on a real task of theirs and leave with it done.
75 minutes
An open hour. Bring a task, leave with a result.
Drop in, optional
8 to 10 minutes in a meeting you already have, plus a one-page handout.
10 minutes plus reading
90 minutes for one team: a shared test question, then each person on their own task.
90 minutes
Pairs do one task twice, by hand and with AI, and compare in a short group session.
About 2 hours over two weeks
The exercises work in any AI assistant, including European and open-source options, so the training follows the tool your organisation chooses.
Recent work
A team of six whose views on AI ranged from ethical refusal to building their own tools. Two afternoons in person opened with a blind read: two versions of the same impact report, one written by people and one by AI, judged before the reveal. From there the team explored possible futures, named what it would not hand over, and each person designed one small experiment for their own work.
Online sessions over the following months tested those experiments on real tasks, with particular attention to what AI summaries lose from local and non-Western sources.
For 2027 the organisation asked to take the approach further: advice on its AI guidelines, a session for all staff, and coaching for colleagues across teams.

After a programme, a fixed monthly package of talks, coaching and clinics, adjusted every three months to what people actually use. Programmes and one-off workshops are priced after a first conversation.