The method
Direct. Cast. Rehearse. Interpret.
Most change gets staged like a product launch: finish the plan, announce it, train people, and wait for adoption. Theater does not work that way.
A director does not hand out the script and hope. The work is four moves, and they happen in order. We call the practice TheaterThink®—the applied method behind everything on this site.
Move 1
Direct: What play are we in?
Name what matters before asking people or AI to act.
The core move. Define the point, audience, stakes, boundaries and evidence of a good result.
Move 2
Cast: Who and what belong in the room?
Compose the ensemble the work needs.
The core move. Bring together people, AI, data, tools, permissions and different perspectives.
Move 3
Rehearse: What happens when the work moves?
Try the workflow while the cost of being wrong is low.
The core move. Run a normal scene and a pressure scene, including ambiguity, exceptions and failure.
Move 4
Interpret: What did the attempt teach us?
Read the evidence and decide what deserves to happen next.
The core move. Notice consequences and choose: Pilot, Revise or Stop.
Why rehearsal is the move everyone skips
Organizations buy training and call it rehearsal. They are not the same thing. Training is being told. Rehearsal is trying, failing in a room where failure is free, and adjusting before it matters.
A bounded experiment makes assumptions visible. It produces evidence instead of opinion, and it lets a team disagree about something real. That is why our engagements are built as a series of attempts rather than a single plan handed over at the end.
Rehearsal is a way of knowing.
Where AI fits
AI is part of the work, not the boundary of it. We ask people to become better directors of it. Every response a model gives is generated through a probabilistic process; fluency is not the same as truth. So we treat important outputs as material to evaluate, not authority to obey.
That is a directing problem before it is a tooling problem. Prompting is rewriting the lines. Directing changes what the performer wants.
Work inside the boundaries that matter
A useful rehearsal is bounded. We involve the people who own technology, information security, legal, accessibility, and governance early enough to shape the experiment—not after a prototype has already created a problem for them.
Constraints become part of the direction: what data can be used, which decisions must remain human, what evidence is needed, and what would stop the work from moving forward. That makes experimentation safer and the eventual handoff more believable.
The idea underneath
TheaterThink is the practice. The larger capacity it develops is called Artistic Intelligence: the ability to interpret what matters, compose with what is available, and discover direction through rehearsal. It is the subject of Josh's forthcoming book, and it is bigger than AI—AI is what makes it urgent right now.