MillerKnoll AI
← All case studies
TracE · IT Service Desk· Jul 2026 · 3 min read

The doctrine that shaped every build since

TracE — AISE’s first agent, built and run by non-developers — processed thousands of IT articles never written for retrieval. The hard lessons became the house doctrine.

TracE was AISE’s first-ever agent — built and still run by non-developers — grounded in thousands of pre-existing IT knowledge articles that were never written with AI retrieval in mind. Getting it working meant processing that entire body of documentation at a pace no manual review could match.

Doing that surfaced hard lessons about how AI actually consumes documentation: unstructured prose confuses retrieval and invites the model to fill gaps with a guess; tables and single-topic articles ground answers reliably instead. Those lessons hardened into AISE’s foundational doctrine — deterministic design, structured-knowledge-first, cite-or-abstain.

It’s the same rule that told the Scott AFB build not to let a part number get inferred, and that anchors every Deviations Knowledge lookup in AutoBid. The first agent didn’t just ship — it wrote the standard the portfolio is built on.