A county internal audit team is designing work on an early public-health AI pilot and must choose between assurance and advisory objectives. What planning distinction matters most?

Clarifying whether the engagement is assurance or advisory when designing the civic AI audit works because objectives, independence constraints, criteria, and reporting differ. Early public-health AI pilots often need advisory help building controls, while assurance engagements opine against criteria with stricter independence. Confusing the two can produce reports that look like opinions without sufficient evidence, or advice that residents and boards misread as formal assurance. Planning must lock engagement type before fieldwork design.

Aligning pilot audit scope with enterprise risk assessment and county stakeholder priorities for public health fails because it skips the threshold decision of whether the work is assurance or advisory; without that lock, criteria and independence rules cannot be set correctly. Establishing sampling methods and evidence thresholds for testing county AI model performance across pilot sites fails because sampling design depends on first fixing the engagement type; criteria and evidence thresholds differ between assurance and advisory work. Agreeing reporting timelines and county resource assignments with public health leadership before starting pilot fieldwork fails because logistics should follow the engagement-type decision; sequencing resources before choosing assurance or advisory risks mismatched expectations and independence gaps.

Exam caveat: advisory work on a pilot can later conflict with assurance on the same system; CAEs should plan cooling-off or scope boundaries so independence is preserved when assurance is later required. Operational check: record engagement type (assurance vs advisory), objectives, criteria, and independence considerations in the planning memo for the health AI pilot; obtain stakeholder acknowledgment before testing starts.