← All case studies

Case study 05 of 15

The Best AI Use Case Was Not AI

A team considered AI for stale data identification and cleanup — and found a better answer.

James Staud · Anonymized field pattern — details generalized to protect specifics

Business challenge

The underlying need was deterministic metadata analysis, reporting, and rules-based archiving, not language understanding.

Architecture approach

Scripting and automation came first, with AI reserved for classification tasks that might genuinely need it later.

How the work moved from request to production

  1. Business need
  2. Intake & risk classification
  3. Select reusable pattern
  4. Build / configure
  5. Review & validate
  6. Deploy / enable
  7. Monitor usage, risk, cost, value

Feeds back into intake to improve the pattern or the governance around it

Governance considerations

Deterministic automation was far easier to explain, test, and approve than a model-based approach would have been.

Results

The team found a faster, lower-risk path to the same value, without a model in the loop.

Lesson learned

AI maturity includes knowing when not to use AI.

Discussion questions

  • Which tasks are actually deterministic?
  • When does classification justify reaching for a model?
  • How do you avoid choosing AI for its own sake?