Map a business process with AI: handoffs, exceptions and accountable ownership

A worked example: where the task becomes difficult

A fictional repair company receives enquiries, estimates jobs, schedules visits and closes invoices. Its staff describe the normal path but omit what happens when a customer changes the address or a technician cannot attend. Begin with one completed job and one failed handoff. A process map should show where information waits, who decides and what evidence indicates completion, not only draw attractive arrows.

Decisions to make before implementation

Distinguish actions, decisions, states and records. Identify which system is authoritative for each fact and which team owns an exception. Define service expectations where known, but do not invent response times. Mark manual work deliberately; it may be a necessary approval rather than inefficiency. Automation should target a verified bottleneck with a recoverable failure path.

A practical sequence for the work

Use the sequence below as a task boundary, not as a claim that the example has been executed. Work with approved inputs and the project’s actual architecture. If a required integration or permission is unavailable, keep that stage visibly incomplete rather than generating a plausible substitute result.

• Walk through one real permitted case and record each state transition.
• Add an exceptional case and identify the owner of every handoff.
• Ask AI to draft a map with waiting points and unresolved rules.
• Verify the map with operators before selecting one automation opportunity.

A detailed brief you can adapt for your agent

Replace the illustrative context with your approved facts and controlled inputs. Keep the stated boundaries when adapting the brief. The expected deliverable matters more than a particular tool name: ask for an explanation grounded in the inspected material and evidence for the requested outcome.

Map this repair workflow from the permitted case notes. Separate actions, decisions, waiting and record changes. Include address corrections and technician cancellation. Name owners only where supported and flag missing rules. Suggest one bounded automation opportunity, but do not send messages, place orders or modify operational systems.

Failure modes that an attractive preview can hide

A model can simplify away the very exception that makes the process difficult. Keep cancellation, missing information and escalation visible. Do not grant an automated agent purchasing or customer-communication authority just because an arrow names that action. Confirm the operational authorization and integration contract before turning the map into executable work.

Technical references: NIST: AI Risk Management Framework

Acceptance checks and the evidence to retain

Keep the owner-verified process map, exceptions and unanswered rules. The selected automation brief should include trigger, permitted action, success evidence and failure owner.

Specific answers

Common questions