# Access AI for Access Reviews {% hint style="info" %} **Early Access:** This feature is available for Veza SaaS customers. It requires Access AI to be enabled for your tenant and the `UI_ACCESS_AI_AWF_ENABLED` feature flag. Contact Veza support to activate these optional capabilities in your environment. {% endhint %} ### Overview Veza Access AI can provide insights into your organization's Access Reviews, helping different stakeholders understand review context through natural-language summaries. These explanations are generated from the current review, as well as information from the parent review configuration. **Access reviewers** can use these insights to quickly understand their assigned work, including the scope of the review, what work remains, when their work is due, and why they received the assignment. **Review administrators and operators** can use insights to verify review scope, assess progress across assignees, and communicate review context to stakeholders without manually interpreting query logic. The review insights panel generates personalized summaries based on: * The types of access and entities under review * Current progress (pending, approved, rejected, and fixed counts) * Review deadlines and schedule information This contextual information helps everyone involved in the access review process understand *why* a review exists, *what* needs attention, and *when* action is required. ### Prerequisites To use Access AI explainability for access reviews, your Veza tenant must have: * **Feature flag enabled**: `UI_ACCESS_AI_AWF_ENABLED` (contact Veza support) * **Access AI** must be enabled for your tenant **Access Controls**: Access AI can only analyze information that the current user has access to via their current role. * Users can generate and view insights for all reviews where they are assigned as a reviewer. This includes reviewers, administrators, operators, monitors, reassigners, watchers, viewers, and NHI security administrators. * Users with the Access Reviewer role will see an explanation of their assigned rows – the summary will not include information about all rows in the review. ### Accessing review insights To open AI-powered insights for a review: 1. Navigate to **Access Reviews** and open a review. 2. In the reviewer interface, locate the **Access AI** button in the toolbar above the review table. 3. Click the **Access AI** button to open the Access AI side panel. ![Access AI button in the reviewer interface toolbar](/files/OLOLOOWhsoIoAxpQjXwK) 4. In the **Suggested Actions** section, click **Explain this Review.** ![Explain this Review option in the Suggested Actions section](/files/YCv5lYYdHIrIjd5iisQF) Explanations appear in a sidebar on the right side of the review interface, displaying: * A formatted summary generated by Access AI * Key metrics about your assigned rows * Information about the review's scope, schedule, and due dates #### When summaries are updated Access AI generates a fresh summary each time you open the Access AI panel. If an error appears when loading the explanation: 1. Close the error message to retry loading the summary. 2. If the error persists, close and reopen the panel, or contact the Veza support team for assistance. Summaries reflect the current state of the review at the time you open the panel, including recent changes to assignee progress, review metadata, or query configuration. ### Understanding the AI summary ![Access AI sidebar showing review insights with review context, status counts, due date, and schedule information](/files/eStN8Lg5brTQzvxMYaDp) The Access AI side panel presents technical information about the review in a structured, readable format. This includes an explanation of the review and configuration metadata, including: **Review Context**: Basic information about the review including its name, due date, and the assignee's role. **Key Statistics**, including current counts for: * Total assigned rows * Pending review items * Items pending sign-off * Accepted and rejected items * Fixed items (if applicable) **Query Summary**: A human-readable explanation of what the review query analyzes, including: * Source entities (e.g., local users, service accounts) * Destination resources (e.g., applications, cloud resources) * Specific permissions or access types under review * Any special filters or conditions **Schedule Information**: For scheduled reviews, details about: * Recurrence frequency * Next scheduled run * Schedule configuration ### How insights are generated When you open the insights panel: 1. **Veza retrieves review metadata,** including the configuration, query specification, assignee statistics, and schedule. 2. **Access AI analyzes the metadata** using an industry-leading LLM model 3. **A personalized summary** is generated and displayed in the panel. Summaries are generated on demand to reflect the current state of your review. Initial generation may take several seconds, depending on complexity. Review data is never sent to public AI services or third-party providers outside your Veza environment. Only review metadata is processed, not the actual review contents (such as row-level information). ### Limitations and known issues {% hint style="info" %} **Note:** This feature is in active development. Limitations and capabilities will evolve in future releases. {% endhint %} Some limitations currently apply when working with Access AI explanations: * Access AI-generated summaries are contextual interpretations. Always refer to the actual review data for authoritative information. * Summaries are currently generated in English only. * Summary generation typically takes several seconds. Complex reviews with extensive metadata may take more time to analyze. * If Access AI services are temporarily unavailable, the panel will display an error. Reload the sidebar to retry. ### Related topics * [Using the Reviewer Interface](/4yItIzMvkpAvMVFAamTf/features/access-reviews/how-to/reviewer-interface.md) * [Access AI Overview](/4yItIzMvkpAvMVFAamTf/features/access-ai.md) * [Access Reviews Query Builder](/4yItIzMvkpAvMVFAamTf/features/access-reviews/configuration/access-reviews-query-builder.md) --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://docs.veza.com/4yItIzMvkpAvMVFAamTf/features/access-reviews/how-to/access-ai-explainability.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.