You spent weeks writing the manual. Then a supplier changed, a step got reordered, two staff found a better way — and none of it went back into the document. Six months later the SOP describes a store that no longer exists, and new staff learn by shadowing instead.
Why this keeps happening.
Procedures change on the floor, but updating the document is desk work someone must remember to do after the rush. The gap between how work is done and what the document says only grows — until the document loses authority entirely and everything is tribal knowledge again.
The way of working.
Incidents feed the procedure
Complaints, exceptions, and floor changes get captured as short notes in the operations project — by voice, by chat, by whoever saw it. Capture is seconds, not paperwork.
The Agent proposes the revision
Periodically an Agent reads the accumulated incidents against the current SOP and drafts the update — reordered steps, new exceptions, removed dead rules. The manager confirms; the document stays authoritative.
History shows why every rule exists
Each revision is on the record with its reason. When someone asks "why do we do it this way?", the answer is in the document's history, not in whoever has worked there longest.
Our own operational knowledge runs the same loop: symptoms and incidents get filed as cases, and standing documents are revised from them by Agents with a person approving. A store's SOP is the same pattern with different incidents.
Start here.
Put one procedure into a project, start capturing incidents against it, and let an Agent draft the first revision — the core workflow shows the propose-and-confirm loop.