---
id: "concept-risk-free-adoption"
type: "concept"
source_timestamps: ["§ How the Company Gained Buy-In", "§ Balancing Control and Accountability"]
tags: ["performance-management", "psychological-safety"]
related: ["concept-span-of-control-vs-accountability", "action-restructure-evaluations", "quote-safe-harbor-compliance", "contrarian-reward-compliance-over-outcomes"]
definition: "Restructuring performance metrics so employees are not penalized for poor outcomes if they followed AI recommendations, thereby removing the fear of adoption."
sources: ["adoption"]
sourceVaultSlug: "hbr-seg-adoption"
originDay: 9
articleStem: "hbr-edu-41-french-spirits-employee-buy-in"
sourceUrl: "https://hbr.org/2025/12/how-a-french-spirits-company-created-employee-buy-in-for-ai"
sourceTitle: "How a French Spirits Company Created Employee Buy-In for AI"
---
# Risk-Free AI Adoption Incentives

To overcome the fear of relying on an unproven AI system, organizations must restructure performance evaluations to remove the career risk associated with the tool's potential failure. Pernod Ricard operationalized this by creating a *safe harbor* for compliance: if a sales representative followed the AI's recommendations but missed their quota, they were not penalized. Conversely, if they ignored the AI and missed their quota, they faced scrutiny (see [[quote-safe-harbor-compliance]]).

This asymmetric risk profile made adopting the AI the safest professional choice for the employee — effectively transferring the risk of the AI's performance from the individual contributor to the organization. It is the accountability adjustment described in [[concept-span-of-control-vs-accountability]], enacted through [[action-restructure-evaluations]].

**Enrichment note.** HBS Working Knowledge documents this safe-harbor policy in detail, with [[entity-iavor-bojinov]] quoted in identical terms. The HBS podcast frames it as part of building *trust in the development team and system* — employees must feel that developers have their best interests at heart. Conceptually it is a practical instantiation of Amy Edmondson's *psychological safety:* removing punitive consequences for experimentation increases willingness to adopt novel tools, especially early in deployment when performance variance is high. See the contrarian framing in [[contrarian-reward-compliance-over-outcomes]]. A live tension: [[question-long-term-accountability]] asks how evaluation should re-blend results and appropriate-use once the tool becomes the mandatory baseline.


## Related across articles
- [[claim-input-metrics-punish-efficiency]]
- [[contrarian-metric-penalties]]