Turning blunt manager notes into fair, specific review language
Managers hand HR rushed, vague, or harsh performance notes and expect polished, defensible write-ups. AI is good at reshaping raw notes into specific, behavior-based language. But this is exactly where people paste real names, ratings, and disciplinary details they shouldn't — and where the temptation to let AI "decide" the rating or the discipline is most dangerous.
You are a writing coach for managers. Rewrite my rough performance-feedback notes into clear, specific, behavior-based language for a review. Keep my assessment; improve the wording only. Role level: {{role_level}} My rough notes (behaviors and examples, names replaced with placeholders): {{rough_notes}} Review period and the rating I have already decided: {{period_rating}} For each point, produce: the behavior (observable and specific), the impact, and a forward-looking suggestion. Use "situation - behavior - impact" framing. Rules: - Do not change my rating or invent accomplishments, incidents, or metrics I didn't mention. If a claim is vague ("bad attitude"), rewrite it as a request for a specific example rather than stating it as fact. - Keep it professional and non-judgmental about the person; focus on work behavior, not personality or protected characteristics. - Remove language that could read as discriminatory or retaliatory, and flag what you removed. - Do not recommend or decide any disciplinary action, promotion, or termination — that is my decision with HR. Balance developmental and positive feedback.
Fill in your details and the prompt updates live — then copy.
Situation - behavior - impact: - During the Q2 launch, [EMPLOYEE] delivered the assigned reports after the deadline on two of three occasions, which pushed the review meeting back and required rework by the team. Suggestion: Agree on interim check-in dates for multi-part deliverables next quarter. Reframed: "Bad attitude lately" isn't specific enough to include — replace with a concrete example (what was said or done, and when). Flag: Removed "isn't a culture fit" — this phrasing can mask bias; describe the specific behavior instead.
The full workflow
- Replace real names with placeholders and strip ratings tied to identifiable people before pasting
- Run the prompt, then reinsert names and specifics in your own secure HR system
- Confirm every rewritten point still matches what actually happened — don't let polish outrun the facts
- Make the rating and any performance-action decision yourself, with HR, and keep documentation in your HRIS
Watch out for
AI cannot make or justify a discipline, promotion, or termination decision. Those require human judgment, due process, and often legal review, and 'the AI recommended it' is not a defense. Use AI for wording only; the decision and its documentation are yours.
Performance notes are among the most sensitive records you handle. Never paste real employee names, ratings, salaries, health information, or disciplinary details into a consumer AI account that retains and may train on inputs — use placeholders or an approved enterprise tool.
Watch for biased or retaliatory language. AI polish can hide it or introduce it; every characterization must trace back to a specific, job-related behavior.
Where this comes from
Every use case on this site is grounded in real reports from working hr managers — not invented by us.