Performance Review Analyst · Use case
How can we use an AI agent to draft performance reviews for managers to edit on our own infrastructure?
A performance review analyst gathers the evidence for a review period: goals, feedback from peers and the records you choose to connect. It drafts a review that cites each piece of evidence, for the manager to edit. The manager writes the assessment and any rating. The agent runs on your own infrastructure, so feedback and reviews are not sent to a public AI service.
Reviews written from memory.
Of the last few weeks.
- ✗ 01Managers write reviews from memory, so the last few weeks count for more than the year.
- ✗ 02Feedback and reviews are among the most sensitive records a company holds.
Where it stops for a person
Always: the manager decides the assessment.Always. The agent gives no rating and makes no recommendation on pay or promotion. The manager decides, and can see the source of every statement.
It gathers the evidence.
The manager writes the assessment.
Step · 01
Goals, feedback and the records you choose are connected.
Step · 02
A manager asks for a draft for one employee and one period.
Step · 03
The agent gathers the evidence for the period.
Step · 04
It drafts the review and cites each piece of evidence.
Step · 05
The manager edits it, adds the assessment and owns the result.
Reads from · connectors
BambooHR
SharePoint
Confluence
The agent connects to BambooHR for goals and review records, and to SharePoint or Confluence for feedback and the work records you choose to connect.
Acts through · MCP servers
Nothing. It drafts.
What you set up
- Connect the goals, feedback and records.
- Provide the review template and state which records may be used.
- Have HR check a set of drafts for fairness before release.
Where it runs
The agent, its knowledge store and the model run on infrastructure you own or control. You choose the model and serve it yourself, so questions, documents and answers are not sent to a public AI service.
How you get it today
This page is labelled Use case: the agent is built in Katonic Studio. You describe the job, connect what is named above, add test questions and release the agent once it passes them. When the Catalog holds a tested template for this job, this page will carry the Template label and show how to take it.
Does the agent rate employees?
No. It gathers evidence and drafts. The manager writes the assessment and any rating.
Who can see a draft?
The manager who asked for it and the people you name. Access is set by role.
Where are feedback and reviews processed?
On your own infrastructure, with a model you serve. Nothing is sent to a public AI service.