Which business processes should you automate first?
Automate first the work that is frequent, time-consuming, done by several people and touches few systems with little judgement — high impact for low effort.
By Logic Intelligence® · Published October 23, 2026
Five criteria
- Frequency — daily and weekly work pays back faster than quarterly work.
- Time per run — minutes add up: 30 minutes a day is over 100 hours a year for one person.
- People involved — the same task done by several people multiplies the cost.
- Systems touched — every extra system adds integration effort.
- Judgement required — rule-based work is easier to automate than work that needs expert decisions.
How to put a cost on manual work
Hours per year ≈ minutes per run × runs per year × people ÷ 60. Multiply by a loaded hourly cost for the people doing the work (salary plus benefits and overhead). For example, 30 minutes × 250 working days × 2 people ÷ 60 = 250 hours a year. This is a general method; Data Circuit's own scoring model is not published.
Common first candidates
Reports assembled by hand, data re-keyed between systems, approvals chased by email, reconciliations, status updates to customers or suppliers, and month-end checklists.
What to leave for later
Rare, exception-heavy work that needs specialist judgement across many systems usually costs more to automate than it saves — on an impact/effort matrix it sits in Defer.
Fix it or automate it?
If a process is broken or about to change, simplify it first; automating it locks the waste in.
Rank your own workflows in minutes
Data Circuit applies these criteria consistently, places every workflow on an impact/effort matrix and builds a month-by-month AI Roadmap. Your first AI Roadmap is free. See how Data Circuit prioritises workflows or read what an AI Roadmap is. When you're ready to build, Data Arcade's Power Platform Automation & Custom Apps package covers many first automations.
Frequently asked questions
Should we start with the biggest process?
Not necessarily. Start where impact is high and effort is low; the biggest processes often touch many systems and need more judgement.
How many processes should we automate at once?
Start with the top few high-impact, low-effort workflows and sequence the rest month by month.
Do we need AI to automate these?
Often not. Many first candidates are rules-based automation or integration; AI helps where the work needs judgement over unstructured information.