Start with business priorities
An audit should begin with what leadership is trying to change: growth, capacity, margin, service, speed, risk or operating visibility.
This creates a filter for AI opportunities instead of allowing every interesting tool to become a project.
Map workflows, systems and current AI use
Review where repeatable work happens, how information moves, what systems are involved and where teams are already experimenting with AI.
The useful unit of analysis is usually the workflow, not the department or the software product.
Turn findings into a prioritised roadmap
Each opportunity should be scored using value, frequency, feasibility, data readiness, risk and how quickly success can be measured.
The output should include business cases, success measures, technology considerations, governance requirements and the recommended first wave of implementation.