> ## Documentation Index
> Fetch the complete documentation index at: https://docs.elevate.help/llms.txt
> Use this file to discover all available pages before exploring further.

# AI Assist

> How People Protocol uses AI to surface context, recommend evidence, and support managers during reviews—without replacing manager judgment.

People Protocol uses **AI** to help managers run performance reviews with better context and less bias.

## What AI does in the platform

* **Surface context** — Summarise activity and artifacts from [connected tools](/platform/connectors-and-evidence)
* **Recommend or populate evidence** — Suggest links between work and objectives or [performance dimensions](/framework/performance-dimensions)
* **Support review completion** — Help structure and phrase feedback; prompt with [checklist](/framework/checklists) items and [evidence](/platform/connectors-and-evidence)
* **Reduce admin** — Less time gathering examples; more time on evaluation and conversation

## What AI does not do

* **AI does not decide [grades](/framework/performance-grades)** — [Managers](/process/roles-and-ownership#managers) complete [checklists](/framework/checklists) and [scorecards](/framework/scorecards); [calibration](/process/calibration) and [leadership](/process/roles-and-ownership#ceo--top-management) approve final grades
* **AI does not replace [evidence](/platform/connectors-and-evidence)** — It surfaces and connects evidence; it does not invent it
* **AI does not override manager judgment** — It assists; the [manager](/process/roles-and-ownership#managers) remains accountable

## How it supports consistency and bias reduction

* **Same structure for everyone** — AI works within [Delivery, Skills, Behaviors](/framework/performance-dimensions) and [checklists](/framework/checklists), so evaluations stay comparable
* **Evidence in context** — Reduces reliance on memory and recency
* **Clear, observable prompts** — Keeps reviews aligned to the [framework](/framework/overview) and [Talent Bar](/framework/talent-bar)

## In the workflow

* **Objective setting** — AI can recommend or pre-fill from [connected systems](/platform/connectors-and-evidence) (see [How it works](/platform/how-it-works) → Define expectations)
* **[Performance reviews](/process/quarterly-cycle)** — AI helps link [evidence](/platform/connectors-and-evidence) to [dimensions](/framework/performance-dimensions) and suggests phrasing
* **Manager check-ins** — [Managers](/process/roles-and-ownership#managers) can surface context and history when needed

AI assist is part of how the platform makes [evidence-based](/platform/overview), [checklist-driven](/framework/checklists) reviews practical at scale.