Performance Review Prompt Builder

Build prompts for generating structured employee performance reviews

Takes an employee role, objectives, and feedback notes and generates an LLM prompt that produces a structured performance review with competency sections, your chosen rating scale, and a concrete development plan. It runs free in your browser on Gera Tools, with nothing uploaded.

Last updated Source: Gera Tools

Will the LLM invent achievements?

The generated prompt explicitly tells the model to use your feedback notes as primary evidence and not invent achievements. Quality depends on the evidence you supply, so paste real notes, peer feedback, and metrics rather than leaving it to guess.

Performance review prompt builder

Writing a fair, specific performance review is hard and time-consuming, and a blank page makes it harder. This builder turns your role, objectives, and feedback notes into a structured prompt that produces a balanced, evidence-based review — complete with competency sections, your organization’s rating scale, and a measurable development plan — that you then edit and own.

How it works

You enter the employee’s role and review period, list the objectives or competencies to assess (one per line), paste your evidence, and pick a rating scale. The tool assembles a prompt that instructs the model to produce a fixed structure: summary, per-competency ratings with examples, strengths, development areas, and a development plan. It explicitly tells the model to tie every claim to evidence and not to invent achievements, so the quality tracks the notes you provide. The prompt is built in your browser.

What to include in your evidence notes

The generated prompt tells the model to base everything on evidence you supply. The quality of the draft tracks the quality of your notes almost linearly. Useful evidence categories:

Quantitative outcomes — sales figures, delivery timelines, error rates, performance metrics. These are the strongest evidence because they are concrete and hard to dispute.

Specific examples — particular incidents, projects, or moments where the employee’s behavior was notable in either direction. “Led the Q3 migration that shipped on schedule despite a vendor delay” is far more useful than “performed well on the migration.”

Peer and stakeholder feedback — verbatim or paraphrased comments from colleagues, customers, or cross-functional partners. Including the source (“from the design team lead in their mid-year feedback”) adds credibility.

Self-assessment notes — if the employee completed a self-evaluation, including their key points ensures the review reflects their own perspective, which is particularly important for development areas.

The more specific and varied your evidence, the less the model needs to generalize — and generalization is where AI-generated reviews become interchangeable and unhelpful.

Writing a useful development plan

The development plan section is often the most valuable part of a performance review for the employee, and the most likely to be thin or vague. The prompt asks the model to generate measurable goals, but the model can only be as specific as the development areas you surface. For each development area in your notes, try to note the size and type of the gap (skill knowledge, confidence, habit, exposure) — this context helps the model suggest more targeted development actions.

A useful development goal has three parts: what will be different, by when, and how progress will be measured. For example, “Deliver two external stakeholder presentations by Q3” is far more actionable than “develop presentation skills.”

Tips and notes

  • Feed it real evidence. The output is only as specific as your notes — paste metrics, peer feedback, and concrete examples.
  • Match your existing scale. Pick the rating format your HR system uses so the draft slots straight into your template.
  • You own the result. The model drafts; the manager verifies every claim, adjusts ratings, and signs off.
  • Mind the data. Reviews are sensitive HR records — only paste what your policy allows and prefer no-training data terms.