---
type: "synthesis"
tags: ["evidence-quality", "vendor-bias", "methodology", "meta-caution"]
articles: ["a038", "a040", "a042", "a052"]
id: "cross-evidence-quality-caution"
sources: ["adoption"]
sourceVaultSlug: "hbr-seg-adoption"
originDay: 9
articleStem: "hbr-seg-adoption"
sourceUrl: "(unified vault: 11 sources)"
sourceTitle: "HBR — People Ⅲ-A · Adoption / trust / literacy / psych-safety"
---
A cross-corpus methodological pattern every downstream agent must hold: the striking statistics that make these articles quotable overwhelmingly come from *consultancies, HR-tech vendors, and proprietary datasets* — the directions are robust, the exact magnitudes are not independent.

- **BCG** is the dominant data house, cited by both A042 ([[entity-bcg-d42]]) and A052 ([[entity-bcg-d52]]) for the adoption gap — meaning the corpus's headline gap statistics substantially rest on one vendor's survey program.
- **A040 (Deloitte)** — [[entity-deloitte-d9]] both *measures* trust (TrustID) and *sells* trust-improvement services; figures like the 89% agentic-trust drop ([[claim-trust-drop-agentic]]), '144% higher trust,' and '10×' are TrustID-specific cuts, not industry facts.
- **A042** leans on vendor self-report: the sabotage stats from [[entity-writer|Writer]] (an enterprise-AI vendor) and the depression link ([[claim-ai-increases-depression]]) are the weakest-verified in the corpus.
- **A038 (BetterUp/Stanford)** — [[entity-betterup-labs]] datasets behind the 61% ([[claim-trust-reduces-workslop]]) and 2–6% figures are proprietary; direction supported, exact figures not.

Synthesis: treat this corpus as a **research-informed argument built on practitioner data**, not settled science. When citing any percentage, name the source organization and flag vendor incentive/self-report where relevant. A parallel Deloitte life-sciences study reporting *different* magnitudes is the cleanest proof that these numbers move by segment. This caution is essential to fairly evaluating the adoption gap ([[cross-adoption-perception-gap]]), the resistance spectrum ([[cross-resistance-spectrum]]), and the measurement debate ([[cross-measurement-problem]]).