← All discussions

P
Leadership & Management

Boardroom pushback on AI generated performance reviews

Penelope Graham · 5 Nov 2026

After the UK government announced new AI transparency guidelines in June, many firms are rolling out AI tools for performance management. My company just piloted an AI system that drafts review comments, but senior managers are flagging bias worries and legal exposure. How are you balancing the speed gains with the risk of employee pushback and potential compliance issues?
👍 395 💬 3

3 replies

F

Finn Sanchez · 5 Nov 2026

I ran a three‑month pilot with SAP SuccessFactors’ AI reviewer set at a 0.8 confidence threshold and every line got a bias flag from senior leads. We scrapped that approach and now we only let the tool tidy up wording after a manager drafts the comment.

David Hernandez · 5 Nov 2026

Honestly we ran a quick test with an IBM Watson pilot last quarter and the bias flags popped up fast. I think the speed gain is real but we had to add a manual review step to keep legal exposure low, which actually slowed things a bit. Overall it’s a trade‑off we’re learning to manage and the data we collect will guide the next rollout.
K

Kathryn Hamilton · 5 Nov 2026

We pushed the AI reviewer from Cornerstone into our quarterly cycle before checking the legal team, and it ended up flagging a lot of gendered language we hadn't anticipated. I realized too late that we needed a simple human sanity check step, so we paused the pilot and added a manual sign off before any comment goes live. Now we're tracking the false positive rate and have set a 0.7 confidence cutoff to keep the pushback down.

Log in or create a free account to join the conversation.