Identifying Automatable Tasks
A framework for auditing your work and finding the highest-value automation targets.
The Automation Opportunity Audit
Not every task should be automated. The best automation targets are: high volume, low variance, rule-based, and currently time-consuming. A simple framework for finding them:
Volume — How often does this task happen? Daily beats weekly beats monthly.
Variance — How different is each instance from the last? Low variance tasks are easier to automate reliably.
Time cost — How long does this take manually? Automate the slow ones first.
Error cost — What happens when this task is done wrong? Automate tasks where errors are expensive but detectable.
The Automation Readiness Score
Score each candidate task on 1-5 scale across four dimensions:
- How often does it occur? (1=rarely, 5=many times/day)
- How structured is the input? (1=freeform, 5=standardized)
- How verifiable is the output? (1=judgment call, 5=objective criteria)
- How painful is it currently? (1=easy, 5=major bottleneck)
Total > 16: automate now. 12-16: automate next. <12: manual for now.
Common High-Value Targets
- Data transformation — converting between formats, normalizing data
- Report generation — weekly status reports, performance summaries
- Email triage — categorizing and routing inbound email
- Content creation — first drafts, social posts, product descriptions
- Code review — automated pre-review before human review
- Research — competitive monitoring, news digests, lead qualification
The 80/20 Rule for Automation
Automate the 80% case first. The remaining 20% of edge cases will consume 80% of the automation engineering effort — often not worth it. Build for the common case; handle exceptions manually until volume justifies further automation.