Task Fit & Everyday Use
Augmentation vs automation
Decide whether AI should assist a person or run a process with little or no human involvement.
5 min readTask fit
Workplace example
Sensitive decision support
AI can generate options for a manager to consider before a sensitive decision. It should not issue the final disciplinary decision or approve an exception without human judgement and policy review.
What this means
- •Augmentation means AI supports a person who still reviews, edits, decides, or approves.
- •Automation means the system carries out more of the process itself, sometimes including classification, routing, or decisions.
- •Sensitive or high-impact work usually belongs in augmentation first, with clear human review before anything scales.
Why it matters
- •Full automation can create speed while hiding errors, bias, or weak assumptions.
- •Augmentation lets teams learn where AI is useful without removing judgement too early.
- •The right level of AI support depends on risk, reversibility, sensitivity, and evidence quality.
Common mistakes
- •Treating every repeatable task as safe to automate.
- •Removing review just because the AI explanation sounds reasonable.
- •Using AI to make employee-affecting or customer-affecting decisions without oversight.
What good judgement looks like
- •Use augmentation when impact is high or facts are uncertain.
- •Use automation only when the task is bounded, tested, reversible, and governed.
- •Keep a person accountable for exceptions and edge cases.
Try this at work
- •Pick one workflow and mark each step as human-led, AI-assisted, or automated.
- •Identify where a human review point must remain.
- •Name what could go wrong if that review point disappeared.
How this helps your reassessment
- •You can tell when AI should assist rather than decide.
- •You consider impact and reversibility before automation.
- •You can define human review points for AI-shaped work.