TracE
MillerKnoll’s Service Desk Copilot — AISE’s first production agent, and the doctrine every later build inherited.
The story, in brief.
TracE is MillerKnoll’s IT Service Desk Copilot and the first agent AI Solutions Engineering ever put into production — built and still run by non-developers. Getting it working meant teaching AI to read thousands of knowledge articles that were never written for retrieval, and the hard lessons became the house doctrine: deterministic, structured-knowledge-first, cite-or-abstain.
How it’s built.
The first build, then called the Service Desk Copilot. Many Confluence docs were outdated, so the team exported a controlled, vetted set of knowledge as PDFs — 100+ IT knowledge-base articles uploaded in 17 batches — and grounded a single agent on those. A deliberate temporary measure to prove technicians could get accurate, cited answers from our own documentation.
Steps through 3 generations — from Stage 0 to Stage 2.
What AI actually does here.
Reading the knowledge base at scale
Processed a body of IT documentation no manual review could keep pace with — surfacing what was outdated, duplicated, or unstructured, and restructuring 2,000+ articles into the retrieval-friendly, single-topic standard.
Curated a vetted POC corpus
While Confluence was being cleaned up, exported and vetted 100+ IT knowledge-base articles as a controlled corpus so the POC could prove accuracy on trusted content first.
Glossary & abbreviation generation
Authored glossaries and abbreviation guides across three Confluence spaces so retrieval resolves the acronyms and shorthand technicians actually type — a common failure point for naive RAG.
Built with the business, not for it.
TracE is notable for who runs it: it was built and is still operated by non-developers on the IT Service Desk, with AISE providing the platform and doctrine.