How to run the audit (90 minutes)
Block time with someone who does the work — not only the manager. Fill lists honestly. Score candidates. Pick one pilot, not twelve.
1. Inventory
- List recurring tasks done weekly or more (even “small” 15-minute ones).
- Note tools touched (email, CRM, sheets, Slack, billing, project tools).
- Mark handoffs between people or systems — handoffs are automation gold.
- Flag rework caused by missing data or inconsistent formats.
- Note customer-visible work vs internal-only work (risk differs).
2. Score each workflow (1–5)
- Frequency — how often it runs
- Time cost — minutes per run × frequency
- Error cost — money, reputation, or delay when wrong
- Rules clarity — can a careful junior follow a written SOP?
- Tool access — APIs or exports available?
High frequency + high time + high rules clarity → automation candidate. High language/judgment needs → agent candidate. High error cost → require human approval gates.
3. Discovery questions
- What starts this workflow (trigger)?
- What does “done” look like?
- What exceptions happen monthly?
- Who is accountable when it fails?
- What report would prove a pilot worked?
4. Choose a pilot
One workflow. Fixed success metric (hours saved, error rate, cycle time). Fixed timeline. Written human checkpoints. Then scope with this guide.
Want a guided Map session? Email hello@oopa.ai with your top three scored workflows.
Want this applied to your stack?
Fixed-scope pilots for AI agents and automations. Map first. Ship one real workflow. Then run it.