The short answer is: There are some aspects of performance reviews that we can hand over to Artificial Intelligence (AI), and there are some we can’t.

AI has the potential, when used properly, to speed up the mechanical aspects of performance management, allowing leaders to focus on judgment. It can assemble evidence into a clean brief, flag bias patterns and rating anomalies, suggest precise coaching prompts during check-ins, and help managers connect objectives to outcomes and behaviors metrics. This is simply to say that AI can be a tool to summarize, surface, and suggest so that managers spend less time hunting data and more time thinking about feedback and fair ratings. In similar ways, AI can be used by peer reviewers to search artifacts like emails, chats, corporate social media, etc. to compile interactions with colleagues throughout the year.

While AI can write first drafts of performance reviews, it’s imperative that managers and peers personalize their narratives and never rely on AI for ultimate rankings. Auto-generated reviews often ring false because they feel generic. Moreover, there are risks of bias, and too much reliance on “watchful” AI to summarize daily interactions and artifacts can feel like surveillance. Instead, set guardrails. Keep humans in charge of ratings and pay decisions. Limit data to known work systems, and give employees visibility into what’s being used with a simple way to challenge it.

Let us know if this helps you figure out a better system!