An AI and operations audit is a structured review of how work, decisions, and information move through a business. It identifies bottlenecks, control gaps, manual effort, system fragmentation, and suitable automation opportunities before an implementation begins.

Five questions the audit must answer

1. Where does valuable work wait?

Look for queues around approvals, data entry, document collection, scheduling, quotation, reconciliation, and reporting. Waiting time often reveals more than task time.

2. Where is the founder acting as infrastructure?

List decisions, exceptions, relationships, and reports that depend on the founder's memory or presence. Separate genuinely strategic judgement from avoidable operational dependency.

3. Which information cannot be trusted quickly?

Identify duplicate records, inconsistent definitions, delayed reports, and numbers that require explanation before they can be used.

4. Which handoffs create rework?

Examine what moves between sales and delivery, operations and finance, HR and payroll, or customer service and management. Missing context at a handoff produces repeated questions and errors.

5. Where can automation create measurable leverage?

Prioritise high-frequency, rule-based, information-heavy work with clear inputs and outputs. Do not begin with a rare exception simply because it sounds impressive.

What the output should include

  • A current-state map of the priority workflows.
  • A list of bottlenecks, risks, and dependencies.
  • A future-state operating design.
  • A ranked implementation roadmap.
  • Clear distinctions between process fixes, system changes, automation, and AI.
  • Indicative effort, ownership, and measures of success.
A useful audit tells you what not to automate, what must be fixed first, and what outcome will prove the change worked.

Frequently asked question

How long should an initial audit take?

A focused diagnostic can begin in a 90-minute working session when the scope is one business area and the right decision-makers are present. Complex organisations may need deeper interviews, system review, and process observation before a roadmap is reliable.