We want to use AI to conduct annual employee performance reviews and draft feedback. How do we determine where the technology ends and human leadership must step in to protect our culture?

Question

Tyler Smith · Accepted Answer

Performance reviews are the bedrock of team health. Using AI to write them risks stripping out the humanity and creating a cold, corporate environment that alienates your people. To protect your culture, you must draw a hard line using the EOS framework of LMA, which stands for Leading, Managing, and Accountability.

AI can handle the cognitive preparation work. For example, you can feed an AI tool raw data from your weekly Scorecards, quarterly Rocks progress, and documented Core Values check-ins. The AI can then synthesize this data and highlight patterns in performance. This saves your managers hours of administrative work.

However, the actual evaluation and communication must remain completely human. A machine cannot measure a team member's GWC, which means whether they Get It, Want It, and have the Capacity to do the job. GWC is about human connection and conative fit, which requires personal observation.

The feedback delivery must also be human. Your managers must deliver reviews face-to-face, using the data summary as a guide, not as a script. If you delegate the actual coaching or tough conversations to an AI-generated script, you abdicate your role as a leader. Keep the preparation technical, but keep the connection human.