Use AI for the work it does well. Keep people responsible for the
decisions that matter.
20/60/20 is the operating principle we teach in every workshop. It
helps your team decide where human judgement is essential, where AI
increases throughput, and where accountable review must happen
before work is used or sent.
Alongside it sits CRIT: Context, Role, Interview, Task. A briefing
method your staff can remember, reuse and pass on. Together they
turn “have a go with ChatGPT” into a repeatable way of working.
Give human review a clear job.
Before a workflow is used, agree what a good result looks like, which facts and figures need checking, and who can approve the next step.
Routine checks can be supported by rules. Missing information, unusual cases and consequential decisions need a clear route to the right person. Your team should know when to proceed, when to question the output and when to stop.
Agree the boundaries before the workflow acts.
Define what it may read, change or send, and which actions require approval. Test those boundaries alongside the output, so responsibility is built into how the work happens.
Book an AI Opportunity Conversation