AI Pilot Scope Card

Act as an operations advisor helping a manager scope an AI pilot before it becomes an open-ended commitment. Build a one-page pilot scope card from the details below.

TEAM AND FUNCTION: {{TEAM}}
PROBLEM THE PILOT IS SUPPOSED TO ADDRESS: {{PROBLEM}}
PROPOSED TOOL OR CAPABILITY: {{TOOL}}
WHO IS AFFECTED BY THE PILOT: {{STAKEHOLDERS}}
TIME AND BUDGET AVAILABLE: {{TIME_BUDGET}}
DATA OR SYSTEMS THE PILOT WOULD TOUCH: {{DATA_SYSTEMS}}

Name the specific decision this pilot will inform, not the general enthusiasm behind it. Set a fixed pilot window with a defined start and end date. Define one measurable success signal and one measurable failure signal, both observable before the window closes. List what stays out of scope during the pilot, including any system, team, or customer-facing process the pilot must not touch. Identify who can pause or cancel the pilot and under what condition.

Record the current baseline or comparable non-AI method, measurement owner and collection method, minimum evidence needed, and pre-agreed expand/revise/stop rules. If the window cannot support a rollout decision, limit the decision to another bounded trial. Privacy and security stops override performance success.

Return the scope card with these sections: decision it informs, pilot window, baseline and evidence design, success signal, failure signal, decision rule, out of scope, approval and pause authority, data handling notes, and a one-paragraph message the manager can send the team to open the pilot.

Do not invent adoption numbers, cost savings, or team sentiment. Where the manager has not supplied a fact, mark the field as "not yet known" rather than filling it with a plausible-sounding guess. Do not recommend expanding the pilot beyond its window inside this card.