Decide what your review should achieve.
Start by deciding what employees and managers should understand and do by the end of the review.
Use the review to recognise achievements, discuss where someone needs to improve and agree development goals. Include the support the manager or organisation will provide, and when progress will be reviewed.
Check that expectations for each role and level are clear. If you already have a career framework, use it as the shared reference. If you do not, agree the expectations before asking people to assess performance.
Choose one person to coordinate the cycle. Set dates for self-reviews, feedback, assessments and review conversations. Once the process is clear, decide which preparation and coordination tasks AI can support.
Choose where AI will help.
Use AI for clearly defined preparation and coordination tasks. For each task, agree who checks the result and what information the tool can use.
Explain expectations
Answer questions using the agreed career framework or role expectations.
Set the expectations
Leaders and the team agree what good work looks like for each role and level.
Coordinate the cycle
Schedule meetings, send reminders and track which tasks are complete.
Run the process
The review organiser sets dates, assigns responsibilities and follows up on delays.
Prepare review materials
Bring together self-reviews, feedback and goals for the review period.
Review the evidence
Managers check the information and consider the context before writing an assessment.
Prepare for the meeting
Gather the feedback, assessment and questions to discuss.
Have the conversation
Managers explain the assessment, hear the employee’s perspective and answer questions.
Support follow-up
Record agreed goals, send reminders and arrange check-ins.
Support development
Managers and employees agree next steps and the support needed to make progress.
Tell your team how AI will be used.
Explain AI’s role before the cycle begins. Give employees and managers clear answers to these four questions.
What will AI do?
Name the tasks AI will support, such as scheduling feedback conversations or preparing review materials. Explain what remains with people.
What can it access?
Describe the information the tool will use: for example, role expectations, previous goals and feedback. Limit access to what it needs for the agreed tasks.
What will be shared?
Explain which notes or documents are shared, who can see them and when. Base this explanation on your tool’s actual settings.
Who is responsible?
Identify who reviews AI’s output, makes the assessment and answers questions. Make clear that ratings and decisions about pay or promotion remain with people.
Discuss this in a team meeting and keep the explanation in writing. If you use an AI agent, let the team try it, ask questions and understand its limits before the cycle begins.
An example review cycle in seven steps.
Use this seven-step cycle as a starting point. Adapt the roles and timing to your organisation, and keep a person responsible for each stage.
Start the cycle
Give everyone the dates, responsibilities and expectations for the review.
AI can: Answer questions about the timetable, review steps and agreed expectations.
Your team: The cycle organiser sets dates and responsibilities. Leaders explain the expectations and AI’s role.
Self-review
Ask employees to reflect on achievements, challenges and development goals.
AI can: Show previous goals and expectations for the employee’s current and next level. Ask questions that bring out specific examples.
Your team: The employee selects the examples and checks that the self-review reflects their own experience.
Peer feedback
Assign a feedback coordinator to speak with colleagues the employee has worked with. This can be the manager or another team member.
AI can: Arrange feedback conversations, send reminders and track which are complete.
Your team: The employee names relevant colleagues. The coordinator speaks with them and gathers examples.
Manager’s assessment
Bring the employee’s self-review, feedback and goals into one assessment.
AI can: Assemble the materials and organise the feedback into themes.
Your team: The manager reviews the evidence and context, then writes an assessment.
Calibration
Check that people in comparable roles are assessed against the same expectations.
AI can: Organise comparisons by role and level, and flag differences for discussion.
Your team: Managers and leaders discuss any differences and agree the final ratings.
Review conversation
Discuss the assessment and give the employee time to respond.
AI can: Schedule the meeting and prepare the review materials.
Your team: The manager explains the feedback and rating, listens and answers questions.
Goals and check-ins
Turn the review into a development plan with regular follow-up.
AI can: Record the agreed goals, send reminders and schedule check-ins.
Your team: The employee and manager agree actions, support and the next check-in date.
Follow through on development.
The review conversation should lead to a plan the employee and manager can use. Agree which skills or behaviours to work on, what support is needed and when you will discuss progress.
Use check-ins to discuss progress, obstacles and changing priorities. Managers provide coaching or arrange practice and mentorship. AI can track agreed goals and send reminders between meetings.
Questions for your next check-in
What progress have you made?
Discuss specific examples of what has changed.
What is getting in the way?
Discuss obstacles and decide what support would help.
What should happen next?
Agree the next action and when you will check in again.
Get ready for your next cycle.
Name the owner and confirm the dates
Confirm the expectations and person responsible for each stage. Include time for review conversations, development planning and follow-up.
Test your AI setup
Give the tool the agreed guidance and appropriate access. Check that its outputs are accurate and that the people responsible know how to review them.
Brief the team and prepare managers
Explain the process, invite questions and make the guidance easy to find. Confirm that managers are ready to discuss feedback and agree development goals.
Resources for your next cycle
- Free review-cycle skill pack
11 Claude Code skills for a full review cycle.
- Review-season guide
A guide to a fair, consistent review cycle.
- Meet Ron
Our AI talent manager for performance reviews.
- Talk with Pauline
Discuss your next review cycle with us.
Use the time you save to support your team.
After the cycle, ask whether AI reduced the preparation work and gave managers more time for feedback and development. Use what you learn to improve the next cycle.
About this guide
Based on the Polar Bear webinar with Pauline Bertry and Alexey Lobachev, published on 4 September 2026.