Most AI adoption conversations happen at the level of the organization: should we use AI, which tool, what policy. The decision that actually bites happens somewhere else, in front of one specific task, in a redesign session, when someone asks "should the machine do this one?"
This template answers that question with three gates, asked in order.
1. Interpretability. Can you understand and explain how the AI reached the result? If not, the task stays human.
2. Impact criticality. If the AI gets it wrong, does someone or something suffer serious consequences? If yes, the task runs as human plus AI.
3. Standardization. Does the task repeat the same way, with well-defined inputs and outputs? If yes, full delegation. If not, back to human plus AI.
Three zones come out the other end: Human decides, Co-piloting, Full delegation. Each one carries the signals that point to it and the cautions that come with it.
What makes it different: not adopting is a legitimate outcome. Most delegation frameworks assume AI participates and only ask how much. Here the red zone is a real exit, so a team can decide not to use AI on a task and walk away with a recorded reason. That is what turns "we don't use AI here" from a gut feeling into a position you can defend.
Who it is for: process improvement facilitators, Lean and BPM practitioners, operations teams, and anyone running an AI adoption workshop. It also works on plain deterministic automation, not only on models that learn, because the gates do not assume a learning system.
How to use it
Write the candidate task or decision in the box at the top. The more specific, the better.
Walk the three gates in order, with the people who actually do the task. Do not skip ahead.
Stop at the first gate that resolves. Record the zone and the reason.
For yellow and green, define a pilot with a scope and a deadline before anything ships.
One rule that makes it work: the people who operate the task run the gates, not the people who know the technology best. The zone is a judgment about the work, not about the tool.
Developed and field-tested in public sector process mapping workshops. Gate criteria build on Shrestha, Ben-Menahem and von Krogh (2019), California Management Review.
