METHOD

Understand a process before changing it.

The method turns an abstract efficiency problem into a concrete decision sequence.

  1. Choose a real process

    Choose one recurring sequence such as an inquiry, quotation, handoff or invoice.

    Follow that sequence through approvals, repeated data entry, system changes, handoffs by call and missing information.

  2. See the process

    Every process transforms an input into an output. Processing, handoffs, waiting points and required information sit between them.

    Many parts do not make a process impossible to understand. If each step remains traceable, the sequence can be mapped and simplified.

  3. Choose the target

    Should this process become faster, better or more profitable? AI can change the current trade-off only when the underlying process architecture changes.

    The target decision comes before selecting an intervention.

  4. Find the binding constraint

    The loudest complaint is not automatically the step that limits total throughput.

    Accelerating a non-constraint does not improve the whole process.

  5. Select the intervention

    The decision order remains visible and fixed.

    AI follows the process, target, constraint and human decision boundary.

    1. Remove
    2. Simplify
    3. Keep the human decision boundary
    4. Make required knowledge explicit
    5. Only then support or automate with AI

Decision sheet

Four fields for the next concrete process decision.

Process
Target
Constraint
Intervention
Look up the seven work areas