How to prepare for a performance review with AI
Prepare for a performance review with AI by organising real examples, practising clear answers and checking your final notes for accuracy.
DIT Team 7 min read

To prepare for a performance review with AI, gather evidence of your work, organise it into clear examples and practise discussing both achievements and areas for improvement. AI can help you structure your notes and rehearse, but you must check every claim yourself and keep sensitive workplace information private. If you are new to these tools, DIT's free AI 101 class introduces ChatGPT, Claude and Gemini for practical tasks, including how to check their answers.
Start with the purpose of your review
A performance review is a chance to discuss your work, understand feedback and agree on what comes next. Your company's process might include a self-appraisal form, a conversation with your manager, or both. Check the questions, criteria and deadlines your organisation has given you before you begin.
Think of AI as a preparation partner, not someone who knows your work or decides your rating. It can suggest a structure, ask useful questions and show where an example needs more context. It cannot verify an achievement from a vague instruction or judge how your manager will respond.
This is different from preparing a CV or a job interview. Your review is about your actual work during a specific period, the expectations of your current role and a constructive conversation about your development. Focus on evidence and learning, rather than trying to write the most impressive-sounding paragraph.
Gather evidence before opening an AI tool
Set aside some time to look back at your work. Depending on what you are allowed to access and use, your notes might include completed tasks, project updates, feedback you received and goals agreed at an earlier review. Check your organisation's rules before copying any information into an AI service.
For each useful example, note:
- The task or goal, in terms you are permitted to share.
- What you personally did, including any support from others.
- The result you can actually verify, such as a completed deliverable or a documented improvement.
- A challenge you faced and what you learned.
- Any evidence or dates you need to check before the conversation.
You do not need a number for every achievement. A specific outcome, such as completing a handover or resolving a recurring issue, can be useful if it is true and relevant. Do not ask AI to invent a percentage, a deadline or a result just to make the example sound stronger.
For instance, instead of writing “I transformed the entire process”, you might say, “I created a checklist for the monthly handover and updated it after feedback from the team.” If you have an approved record of a measurable result, add it accurately. If you do not, leave it out.
Protect confidential and personal information
Workplace material can contain client details, employee information, internal plans or commercial data. Even a performance review can mention sensitive feedback about another person. Before using an AI tool, follow your employer's policy and check which services and data are approved for your role.
If you are unsure, do not paste internal documents or identifying information into a public AI service. Replace names with neutral labels, remove client and project identifiers, and use a general description of the work. An anonymised description is not automatically safe: if the remaining details could still identify someone or reveal confidential information, keep them out.
You can also practise with a fictional example first. Use the structure you learn to write the real version yourself in an approved place. If a tool has not been given a document, it cannot know what that document says, so do not treat its assumptions as evidence.

Turn your notes into a balanced self-appraisal
Once you have a few verified examples, ask AI to help organise them against the review questions. Give it only information you are authorised to share. Ask it to distinguish what you did from what the team achieved, and to flag missing context rather than filling gaps with guesses.
A useful answer usually covers three things: the goal or situation, your contribution and the result or lesson. Keep each point short enough to discuss naturally. If your role included teamwork, acknowledge it instead of claiming sole credit for a shared outcome.
Balance is important. Include an area you want to improve and a practical next step, without presenting an unverified claim about why something happened. For example: “I want to make project updates clearer. I plan to agree on a brief update format with my manager and ask for feedback after using it.” Adjust this to what actually happened and to what you are ready to do.
Read the draft aloud. If it sounds unlike you, simplify it. If it adds a result you cannot support, remove it or mark it for checking. Your review should sound like an honest account of your work, not a generic advertisement.
Try this now: organise one example
Pick one non-sensitive or fictional work example. Write two or three sentences about the situation, what you did and the outcome you can confirm. Then copy and adapt this prompt in an approved AI tool:
I am preparing for a performance review. Help me organise the example below into a short, honest self-appraisal. Separate the situation, my contribution, the confirmed outcome and what I learned. Do not invent numbers, results, dates or responsibilities. If information is missing, ask me a question or mark it as unconfirmed. Suggest one constructive development question I could discuss with my manager. Here is my example: [paste an approved, non-sensitive example].
Compare the response with your original notes. Highlight every new detail the tool added and remove it unless you can confirm it. Then make one version in your own words and save it in a place approved by your employer.
Practise the conversation, not just the form
A written appraisal is only part of the preparation. You can ask AI to play the role of a manager and pose one question at a time, such as “What are you most proud of?” or “What support would help you improve?” Respond in your own words, then ask for feedback on clarity and whether your answer includes a concrete example.
Do not assume the simulated manager's reaction predicts your real manager's view. Use the exercise to notice whether you are speaking clearly, acknowledging other people's contributions and asking for what you need. You can stop the role-play or change its tone if it becomes unhelpful.
Prepare a few questions for the actual meeting:
- Which outcomes should I prioritise in the next review period?
- Where could I improve, and what would good progress look like?
- What support, feedback or learning would help me reach those goals?
If feedback surprises you, ask for an example and take time to understand it. You do not have to agree immediately or turn a practice script into a rigid speech. After the meeting, record agreed actions and owners in your organisation's approved system.
For more practice with one AI assistant, DIT's Master Claude workshop is a live Zoom session priced at ₹99 plus GST. Choose an approved tool for work and continue to check its suggestions against your own evidence.
Frequently asked questions
Can AI write my whole self-appraisal?
It can help structure and edit notes that you are allowed to share, but it cannot know or verify your work on its own. Check each statement and take responsibility for the final version.
What if I cannot quantify an achievement?
Use a specific, truthful description of the work and its observable outcome instead of inventing a number. If a measurement matters, check an approved record or ask how your organisation assesses it.
Should I paste my manager's feedback into an AI tool?
Only if your employer's policy and the tool's approved use allow you to share it. Otherwise, remove identifying and confidential details, or practise with a fictional example and write the real response yourself.
Want to practise using AI for everyday tasks and learn how to check its answers? Join DIT's free AI 101 class, live on Zoom with no coding needed, to explore ChatGPT, Claude, Gemini and a reusable prompt framework.





