The Anchor Method
AI is the means. Impact is the anchor.
Six stages for putting AI to work in an organisation without losing the people, the process or the evidence. It came out of watching the same failure repeatedly: a capable tool dropped into a process nobody changed, handed to people nobody prepared, producing answers nobody checks.
A working lens, drawn from practice.
These six stages hold up in the rooms he teaches in. They have not been tested against a control group, and no study has yet measured whether teams that use them do better than teams that do not. That work is worth doing and has not been done, so the method is offered as a lens rather than as a proven methodology.
The craft inside it, such as how to write a good instruction or how to check an answer, is standard practice rather than anything he invented. Where the thinking draws on other people’s work, they are credited in the material itself.
Six stages, in order.
The initials spell the name. That is a memory aid for people sitting in a workshop rather than a rigid sequence, because real work loops back constantly.
Assess the landscape
There is more AI news in a week than anyone can act on, so the first skill is triage. What to go deep on, what to note and move past, what to ignore, and what to hand to someone else. Most overwhelm is a filtering problem rather than a capability one.
Narrow to one use case
One recurring task that genuinely costs the team time or grief. Starting where the impact is, rather than where the technology is most interesting, is what separates a programme that sticks from a pilot that quietly stops.
Confidence and capacity first
People need permission to start badly. A team that is frightened of looking foolish will not experiment, and a team with no attention left will not either. Attention is the scarce currency here, not budget.
Human judgment and limits
Name the places AI does not go before anyone asks. Decide what must be verified and by whom. Trust the tool, check the output, and keep a person accountable for everything it touches, because someone will eventually have to answer for it.
Operationalise
Move deliberately from occasional use, to reliable use, to built-in. Different modes suit different work. A sustainable rhythm beats a heroic push, which tends to end the week its champion goes on leave.
Roadmap, resilience and spread
Grow the people who will carry it after you leave. Move from one person coping, to a team that works this way, to an organisation that would survive losing any one of them.
What sits under the six.
Three things hold the stages together, and they are why the training does not feel like a tools demonstration.
Awareness, striving, integrity. People adopt something new when they can see clearly, have something worth working towards, and are not asked to compromise themselves to do it. Take one away and adoption stalls, whatever the tool does.
A decision filter. Before any tool: what problem is this, what would better actually look like, and would we know if it worked? Plenty of good answers to that end with “so, not AI.”
Teaching under pressure. He trained in counselling and psychotherapy. He does not use it clinically; it changed how he teaches. People learning something difficult while already stretched need a different kind of room than people being sold a product.
This is what the training is built on.
The lens underpins the corporate programmes, the startup cohorts, the water organisations in East Africa and the public courses.