Find AI Prompts
Home›Human Resources›Performance Reviews
Human ResourcesPerformance & Development

AI Prompts for Performance Reviews

A performance review is a judgment about work, made from evidence, delivered as a conversation, and consistent with how other people's work was judged. Each part fails independently: reviews written from recent memory, reviews that describe traits instead of behavior, reviews that are surprises, and ratings that vary more by manager than by performance. The fixes are evidence gathered across the period, behavior-based language, no surprises because feedback was continuous, and calibration.

This page absorbs the library's performance-conversation prompt and adds the evidence-based written review and the calibration session. The guardrail applies throughout: the manager and HR decide; the model organizes evidence and drafts.

Before you use these

Have these ready to replace the highlighted [variables]:

These prompts structure evidence and draft documents for performance, promotion and talent decisions. They do not make those decisions. Every rating, ranking or recommendation the model produces is a draft for a manager and HR to review against consistent, job-related criteria; nothing here may score people on protected characteristics or their proxies, and decisions with employment consequences follow your policy and local law.

The prompts

1. Prepare the performance conversation

Best forTalking points for a constructive review conversation: strengths with examples, growth areas as behaviors, goals, and phrasing for the hard parts.
Inputs needed
  • The person's situation: strengths, gaps, examples
How to use itAbsorbed from the library. Give real examples — the talking points are only as specific as the input.
Expected outputAn opening that sets a constructive tone, three strengths with examples, two growth areas framed as behaviors, SMART goals for next quarter, and phrasing for the difficult moments.
Act as an HR business partner. Help me prepare a performance conversation with [role] whose situation is: [strengths, gaps, examples]. Produce: an opening that sets a constructive tone, 3 strengths with specific examples, 2 growth areas framed as behaviors not traits, SMART goals for next quarter, and phrasing for the two hardest moments in the conversation — raising a gap they may not accept, and answering 'why wasn't this raised earlier?'

2. Write the review from the period's evidence

Best forA written review built from objectives, evidence and feedback across the whole period.
Inputs needed
  • Objectives and evidence
  • Feedback gathered
  • Rating scale
How to use itPaste the evidence. The model writes the review by objective and competency with the evidence cited, proposes a rating against the scale's definitions for the manager to confirm, and flags recency and other biases in the material.
Expected outputDraft review with evidence per objective, behavior-based strengths and growth areas, a proposed rating with the definition it matches, bias flags, and the items the manager must confirm.
You are drafting a written performance review for [person, role] for [period], for the manager to edit and confirm.

Objectives and evidence: [per objective: outcome, evidence, dates]
Feedback gathered: [peers, stakeholders, one-to-one notes — with dates]
Rating scale definitions: [paste]

1. Per objective: what was achieved against the measure, the evidence (cited with dates), and the context that affected it (dependencies, scope changes).
2. Strengths (2–3): behaviors with specific examples from across the period, and their effect.
3. Growth areas (1–2): behaviors, not traits; specific examples; the expectation going forward; the support to be offered.
4. Proposed rating: the scale definition the evidence best matches, with the reasoning; note if the evidence could support two ratings and what would decide it.
5. Bias flags in the material: recency (evidence clustered in the last two months), a single source dominating, feedback about style rather than work, anything unrelated to performance — excluded and listed.
6. Items the manager must confirm before this is a review: every rating and every example.

Write in plain, respectful language. The draft supports the manager's judgment; it is not the judgment.

3. Calibrate ratings across the team

Best forA calibration session that tests each proposed rating against the evidence and the scale, for consistency — not a forced distribution.
Inputs needed
  • Proposed ratings with evidence summaries
  • Scale definitions
  • Participants
How to use itPaste the proposed ratings and one-paragraph evidence summaries. The model prepares the session: cases to discuss first, consistency questions, and the record — reporting at the process level.
Expected outputCalibration agenda, cases flagged for discussion with the reason, consistency questions, manager-pattern observations, and the session record format.
Act as an HR business partner preparing a rating calibration session for [team / function].

Proposed ratings: [per person: role, level, proposed rating, evidence summary, manager]
Scale definitions: [paste]
Participants and rules: [who attends; no forced distribution; decisions by evidence]

1. Cases to discuss first: ratings where the evidence summary does not clearly match the scale definition; where similar evidence has different ratings across managers; where the rating changed materially from the prior period without evidence.
2. For each flagged case: the specific consistency question to ask (not a verdict).
3. Manager patterns: any manager whose proposed ratings cluster unusually high or low relative to evidence — raised as a calibration point about the scale, privately.
4. Group patterns: whether any group (team, location, working pattern, tenure band) is rated systematically differently for similar evidence — flagged for review as a consistency and fairness matter.
5. Agenda with time by case; the rule that ratings change only when the evidence, not the argument, supports it.
6. Record: what was discussed, what changed and why, actions for managers before reviews are delivered.

The session calibrates ratings against evidence and definitions. Do not propose ratings for individuals; surface the questions.

Related prompts

Logical next step

After this, most HR teams move on to Performance Objectives and Goal Setting.

Get the free HR & Recruiting AI Starter Kit → Nine of these prompts as an intake → assess → offer hiring workflow with an intake worksheet, delivered by email. See what's inside

All Human Resources prompts · Search the full library