Skip to main content

​
What Is the Access Request Agent?

The Access Request Agent is a beta experience that helps employees request access using natural language in the Lumos AppStore and supported AI assistants. Instead of browsing a catalog and guessing which permission to select, employees can describe what they’re trying to do — for example, “I need Salesforce access to update customer accounts” — and the agent will:
  • Recommend the right app and permission (policy-bound and least-privilege)
  • Ask targeted follow-up questions when the request is ambiguous
  • Draft a higher-quality business justification
  • Generate a request preview that the employee can review before submitting
Important: The Access Request Agent does not grant access automatically. Existing Lumos policies and approval workflows still apply. The employee remains in control and must submit the request. The agent runs on your existing configuration — no new policy setup is required. It leverages whatever is set in your AppStore today. More functionality to control agent conversation behavior on a per-app and per-permission basis will be provided in the future.

​
Access Request Agent in AI Assistants and Bots

In addition to the Employee AppStore experience, organizations can also expose AppStore request workflows to approved AI assistants and developer tools using Lumos MCP support. This allows employees to use natural language access workflows directly from compatible AI clients such as Claude, Claude Code, Cursor, Gemini, Codex, Glean, and other MCP-compatible tools. Employees can use these assistants to:
  • Search available applications and permissions
  • Discover least-privilege access options
  • Draft or submit access requests
  • Check the status of recent requests
Examples:
  • “Find the least-privilege Salesforce access that lets me create reports.”
  • “Request Snowflake access for 30 days for Q3 reporting work.”
  • “Show my recent AppStore access requests.”
All requests still respect existing Lumos policies, approvals, scopes, and authorization controls. To learn more about configuring MCP-compatible AI clients with Lumos, see the Lumos MCP Getting Started Guide

​
Who Is This Beta For?

This beta is designed for organizations that want to:
  • Reduce back-and-forth on access requests caused by unclear justifications
  • Help employees choose the correct permission level without training on entitlement naming
  • Reduce time-to-request and improve request completion rates
  • Enable secure AI-assisted access workflows using approved enterprise AI tools

​
How Employees Use It

​
Step 1: Open the Agent

Employees can access the agent from the Employee AppStore (web) wherever the beta entry point is enabled. Organizations using Lumos MCP can also access supported workflows from compatible AI assistants and developer tools.

​
Step 2: Describe What You Need

Use plain language. Good requests include:
  • The app (if known)
  • What you’re trying to do (task or outcome)
  • Any relevant context (team, project, or urgency)
Examples: “I need Jira access to comment on tickets for the Mobile team.” “I need to view Q1 campaign performance metrics — what tool should I use?” “I need Snowflake access for reporting for the next 30 days.”

​
Step 3: Answer Clarifying Questions (When Needed)

If there are multiple valid permission options, the agent may ask a short follow-up question to narrow scope (for example, view vs. edit).

​
Step 4: Review the Request Preview

Before submission, you’ll see a structured preview showing the app, permission, duration, and justification.

​
Step 5: Submit or Switch to the Standard Form

You can:
  • Submit the request after reviewing it, or
  • Switch to the standard request form at any time