AI and automation
How to identify workflows suitable for automation
Not every repetitive task should be automated. A simple filter helps you pick the right candidates first.
Look for volume plus rule
The best candidates are high-frequency tasks that already follow a describable set of steps with clear inputs and outputs.
Score reversibility
Low-risk, easily reversible actions (drafting, tagging, summarising) can be automated early. Irreversible or financial actions should keep an approval step.
Measure the current cost
Record the time and error rate of the manual process before automating. Without a baseline, you cannot demonstrate improvement later.
Check data quality
Automation amplifies messy inputs. If the source data is incomplete or contradictory, fix that first — otherwise the automation produces fast, reliable mistakes.
Design the review loop
Decide who reviews outputs, how failures are surfaced, and how the automation is paused. A safe review loop is part of the design, not an afterthought.