How to Choose Your First AI Automation Use Case (Step-by-Step)
A 5-criteria scoring framework to pick the one workflow you should automate first — the one that ships in 2 weeks and pays for itself in 30 days.
Who this is for: Founders and operators who know they want to use AI but freeze in front of the blank canvas.
The biggest reason AI automation projects fail isn't technical — it's choosing the wrong first workflow. Pick something too big and you'll abandon it. Too small and nobody cares. Here's the scoring grid Opsacea uses with clients.
The 5 criteria
- Frequency — happens at least weekly, ideally daily.
- Pain — currently steals time from the operator or a senior person.
- Boundedness — has a clear start, a clear end, and lives in 1-3 systems.
- Tolerance — a wrong AI output is annoying, not catastrophic.
- Visibility — the result is obvious to the team within a week.
How to score
Score each candidate workflow from 1-5 on each criterion. Sum them. Anything 20+ is a great first project. Below 15, keep looking — you'll burn out before you finish.
Examples that consistently score high
- Inbound lead classification + routing.
- Meeting note summarization with action items.
- Support ticket triage and tagging.
- Weekly client status report draft.
- Job application screening for clear disqualifiers.
Examples that look great but score low
- 'A chatbot for our website' — unbounded, low frequency per session, hard to measure.
- 'AI that writes our marketing' — high tolerance risk, low boundedness.
- 'Replace our SDR team' — too big, too political, too risky for v1.
Next step
Once you've scored your shortlist, the Opsacea Playbook gives you the implementation template for the top 3 workflow archetypes — intake, triage and drafting.
Keep reading
More from the same cluster — internal links that compound.