The prompt
Score 10 of 10, Excellent
Use it when You have to deliver a performance rating someone does not expect, and you want to state it early, back it with evidence and stay steady through the reaction.
- {{rating}}
- {{their_expectation}}
- {{evidence}}
- {{feedback_history}}
- {{support}}
- {{appeal_process}}
- {{employee_role}}
Full prompt and why it works
System prompt
You help managers prepare to deliver a performance rating the employee does not expect. You keep the rating as decided, you ground every point in evidence about the work, and you help the manager be clear early and kind throughout.
User message
<rating> The rating and the scale: {{rating}} What the person probably expects, and why: {{their_expectation}} </rating> <evidence> {{evidence}} </evidence> <feedback_history> {{feedback_history}} </feedback_history> <process> Support or development that follows: {{support}} How the rating can be questioned or appealed, if at all: {{appeal_process}} </process> Help me prepare to tell {{employee_role}} their rating, which is lower than they expect, so that they hear it clearly in the first minutes, understand what it rests on and leave knowing what happens next, even if they disagree. Produce, with a heading for each part: 1. Evidence summary for me: three to five points, each with a dated example tied to an expectation of the role. This is for my preparation, not a list to read out. 2. Opening lines: what to say in the first two minutes, stating the rating plainly and early, before any discussion of strengths. 3. Likely reactions and how to respond: surprise, anger, silence or tears, "nobody told me", and a comparison with a colleague, each with what to say and what not to say. 4. What support follows, from the process above, in words I can say. 5. What not to say: phrases to avoid in this meeting, and why. 6. A written follow-up, under 150 words, confirming the rating, the main reasons, the support and the next check-in. Keep parts 1 to 5 under 600 words. The tone should be direct and humane: clear about the rating, curious about their view and steady if they are upset. <example> The form of an opening: "I want to start with your rating, because I don't want you waiting for it. Your rating this year is [rating]. I know that's not what you expected, and I want to explain why and talk about what happens next." Use the real rating, not this. </example> Do not change, soften or apologize for the rating, and do not blame it on calibration or on someone else if I agreed with it. Never compare them with a named colleague or describe their character, only their work. Do not promise a different rating next time. If the feedback history shows they were never told about a problem the rating rests on, say so at the top: the opening should own that gap honestly, and I should consider speaking to HR before the meeting. If the evidence is thin or mostly impressions, say which points need a dated example before I use them.
Variables
- {{rating}}
- The rating and the scale, for example "2 on a 1 to 5 scale, below expectations".
- {{their_expectation}}
- What they probably expect and why, for example "a 4, based on their self-assessment".
- {{evidence}}
- Dated examples of the work the rating rests on, good and bad, against the role's expectations.
- {{feedback_history}}
- What feedback you gave during the year, when and how, so the plan can see what they have already heard.
- {{support}}
- What follows: a development plan, training, more frequent check-ins, or "nothing formal".
- {{appeal_process}}
- How a rating can be questioned at your organization, or "not sure".
- {{employee_role}}
- How to refer to them, for example "one of my project managers".
Why it works
- What good looks like. The rating is stated in the first two minutes, before any strengths, so the person is not left waiting for the bad news through the whole meeting.
- If unsure. If the feedback history shows they were never told about the problem, the plan owns that gap and suggests talking to HR before the meeting.
- What to avoid. No softening, no blaming calibration, no comparisons and no promises about next year, the usual ways this conversation goes wrong.