---
type: "synthesis"
tags: ["synthesis", "trust", "psychological-safety", "adoption"]
sources: ["execution"]
id: "cross-trust-execution-substrate"
sourceVaultSlug: "hbr-seg-execution"
originDay: 8
articleStem: "hbr-seg-execution"
sourceUrl: "(unified vault: 7 sources)"
sourceTitle: "HBR — Firm Ⅱ-C · Execution quality — correct execution of AI"
---
## The variable under every adoption failure

Three articles converge on trust as the rate-limiter of AI execution:

- **A060** makes it a leadership dimension: Human Centricity ([[concept-human-centricity]]) and [[quote-trust-speed-limit|'trust sets the speed limit for AI adoption']].
- **A076** makes it the empirical spine: [[claim-trust-predicts-hiding]] — the lowest-trust quartile is ~4x more likely to hide AI use (47% vs 14%), holding after controls. The mediating mechanism is [[prereq-psychological-safety-basics|psychological safety]].
- **A093** operationalizes it culturally: Moody's 'yes, and…' mandate and company-wide buy-in ([[framework-moodys-guiding-principles]]) made 14,000 people willing to experiment.

## Why this is the deepest through-line

The corpus's other themes bottom out in trust. Process redesign ([[cross-task-to-process-translation]]) requires employees to volunteer how they actually work. Scaling ([[cross-scaling-discipline-sunsetting]]) requires disclosure of what works. Avoiding [[cross-augmentation-vs-replacement]] requires believing AI won't get you fired. A076's [[quote-trust-battle-lost]] states the failure mode: 'when hiding from your own employer becomes the prudent career strategy, the organization has already lost the trust battle.' Notice the Edmondson connection: [[prereq-psychological-safety-basics]] (A076) and Human Centricity (A060) are the same construct wearing different clothes. Contrast this with the structural-control instinct in [[cross-governance-vs-psychological-safety]].