Six products, two company names, and a fact base that contradicted itself

live fact conflicts found in the first automated drift report

First drift report (run before the pilot began): two main sites named two different companies; the headline question count read 4,000+ on the live site while the team's official figure was 5,589 — a number appearing on neither site. The third conflict was a missing brand asset discovered during the same scan.

The problem

A health-tech company running six products across a corporate site, product sites, app stores, and exhibition materials. The same fact had different values on each surface — including two different company names on the two main sites and a headline question count that matched neither.

How it worked before

Product facts lived in scattered documents, outdated brochures, and team memory. Every update required manually checking six products across every surface. Drift was discovered by accident, usually right before a deadline.

What we built

A canonical Product Brain: one authoritative reference for every product's figures, pricing, positioning, and brand identity — queryable in plain language, with automated drift detection that flags any surface contradicting the canonical record.

What changed

Fact-checking moved from a manual, error-prone process to an automated system that catches contradictions before buyers see them. The team stopped running around looking for the right logo.

System shape

Inputs

Six products' public surfaces + internal fact documents

Processing

Canonical Product Brain + automated drift detection

Outputs

One authoritative reference + drift reports flagging contradictions

Method

1Map: audit all public surfaces to surface every fact conflict and the root cause
2Structure: design a canonical Product Brain schema covering figures, pricing, positioning, and brand identity
3Implement: build the queryable reference and automated drift-detection pipeline
4Adapt: refine the schema against real drift reports and team feedback

Constraints

Six products with distinct brand identities
Public surfaces only — no internal access required for the pilot
Draft-first approach: nothing changes without CEO sign-off

Our role

Transformation lead and strategic advisor to the CEO

Why this is anonymized

Client identity is withheld under engagement confidentiality. The metrics are real and verifiable in a private reference call.

Have an operational document you want analyzed?

We do not run standard sales calls. Send your most complex operational document and an engineer will return a clear model of what automating it is worth.

Request an operational diagnostic

All work