Deciding what to automate
Which work an AI agent should own and which should stay with a person — the tests we apply in discovery, and the signals that say leave it alone.
The service behind these pieces is workflow automation and AI.
Updated
Pieces on this topic
How do digital employees differ from traditional software automation?
Automation executes steps you defined in advance. A digital employee is given an outcome and works out the steps — which changes what you can hand it.
What should an AI agent own?
Work that is continuous, cross-system and cheap to check. Four tests that separate the workflows worth an agent from the ones that cost you.
When should a workflow stay human?
Five signals that a workflow should not be given to an agent — and why an implementer who never says so is telling you something about themselves.
Assistive AI vs autonomous AI — which do you need?
A copilot makes your team faster; an autonomous agent owns an outcome. They are peers, not rungs on a ladder, and most firms need both.
Working through this right now?
Everything filed under “deciding what to automate” is the general case. Bring the specific one: a principal consultant will tell you which of these pieces applies to it, and which does not — usually the more useful answer.