Skip to main content

How Performance Leader uses AI

The principles behind Compass AI, what each AI feature does and what data it reads, what your firm controls, and who to talk to with questions. Written to be linked from anywhere AI appears in the platform.

Performance Leader uses AI, branded Compass AI, in several parts of the platform. This page explains the principles behind it, what it is used for, and what data each feature draws on. It is written to be linked from anywhere you encounter AI in the product.

Our principles

We do not train on your data. Your firm's content is used to produce a result for your firm. It is not used to train or improve AI models.

We do not retain prompts. The text sent to produce a summary or suggestion is not kept after the result is returned.

People remain accountable. Compass AI summarises, analyses and assists. It does not make decisions about anyone, and it is not a substitute for the judgement of the people running a performance process. The platform says so at the point where summaries are configured:

Compass AI summaries can include errors, omissions, or misinterpretations. Always review generated content before sharing or using it in people decisions.

You can see what it is doing. Every prompt that drives an AI feature to end users is visible to your administrators, who can edit it, override it for a single review cycle, or reset it to our default. Administrators can also see what the AI has actually done in your environment: requests, response times and error rates, broken down by module. Nothing about how AI behaves in your firm is hidden from the people responsible for it.

Your firm decides what is switched on. AI is not all-or-nothing. Each module and add-on has an independent control, and turning one off stops its AI processing while the module carries on working normally.

Where AI is used, and what it reads

Feature

What it does

What it reads

Review summaries

Summarises a completed review, including summaries for managers

The responses in that review: self-assessment, reviewer and contributor answers

360 feedback summaries

Synthesises contributor feedback into themes

Contributor responses for that review

Custom review summaries

Summary types your firm writes for a particular cycle

The responses in that review

Review readiness

Checks a review's design before launch

The review's configuration only: forms, steps, emails and reports

Objective setting assistant

Helps someone write a well-structured objective

The objective being drafted, and your firm's plans where they are available

Objective quality review

Scores an objective for clarity and completeness

The text of that objective

Strategic alignment

Explains how firm plans relate to a role

Firm plans and the individual's objectives

Feedback quality and sentiment

Scores feedback across calibrated dimensions, feeding analytics

The text of the feedback given

Personal feedback summary

A development-focused summary for the person receiving feedback

Feedback that person has received

Check-in and debrief summaries

Summarises a conversation or a project debrief

The responses recorded in it

Direct reports and leadership analytics

Engagement patterns and conversation starters for a team

Engagement, feedback and objectives data for people in that scope

Email campaign content

Drafts the wording of a campaign email

Only the instruction the administrator types. No employee data

Not every feature is switched on in every firm. If something here is unfamiliar, your firm has most likely turned it off or not enabled it.

What your firm controls

Your administrators, not Performance Leader, decide which AI features are available, what the prompts say, and who can see certain outputs. Those controls are covered in Enable or disable Compass AI features and Review and update Compass AI prompts.

A small number of features are configured by Performance Leader at your firm's request rather than from a settings page. The effect is the same; the route is a conversation with your Customer Success Manager.

Data processing and security

Questions about data processing, residency, retention periods and sub-processors are answered in our data protection impact assessment and its addendum, which your firm will hold. That document is the authoritative source on those points.

Questions or concerns

If you are using Performance Leader at your firm and something about an AI feature concerns you, speak to whoever administers the platform there. They can tell you what is enabled, show you the prompt behind a feature, and turn it off if that is the right answer.

Administrators should raise anything they cannot resolve with their Customer Success Manager.

Did this answer your question?