We want to use AI to help our managers prepare for their quarterly conversations and reviews, but we are worried it will feel cold or biased. How do we run this process so it supports our Humbly-Confident core value?

Question

Tyler Smith · Accepted Answer

Using AI to assist with performance reviews can be highly effective, but only if you use it to enhance human feedback rather than replace it. To align this process with your Humbly-Confident core value, you must establish clear guardrails. AI should never write a performance review from scratch. That approach feels cold, lazy, and violates the trust of your team. Instead, use AI as an analytical assistant. Have your managers feed objective, historical performance data into a secure AI environment, including weekly Scorecard metrics, quarterly Rock completion rates, and notes from past one-on-ones. Ask the AI to identify objective patterns, such as consistent dips in productivity during specific months or areas where the employee has shown steady growth. This helps your managers overcome recency bias and prepares them for a more balanced, fact-based conversation. Once the AI generates these insights, the manager must apply their own human judgment, empathy, and context to draft the final review. The resulting conversation must be delivered face-to-face with vulnerability and respect. When sharing feedback, the manager should be transparent about using AI to compile the data, showing that they did the work to ensure the review is fair and comprehensive. This demonstrates humility by admitting that human memory is imperfect, while showing the confidence to lead with objective facts. By keeping the final delivery and relationship human, you use technology to make your evaluations more objective, honest, and impactful.