# Descope Adds Approval and Access Controls for Workplace AI Agents

By Simon Yoon

Canonical URL: https://www.tokenpost.com/news/technology/32650
Published: 2026-10-15T17:40:57.000Z
Updated: 2026-10-15T17:40:57.000Z
Section: Technology

> Agentic Identity Hub 3.0 adds employee review, MCP gateway integrations and new authentication methods as AI use on corporate devices expands.

Descope is adding employee approval and access controls for workplace AI agents, seeking to help organizations govern their use as AI adoption on company devices grows.

Descope announced Agentic Identity Hub 3.0 on Thursday, Oct. 15. The release extends identity controls to AI agents and Model Context Protocol (MCP) servers, which connect agents to tools and services.

The new request process lets employees identify an agent and list the applications, APIs, data sources and MCP servers it needs. A reviewer must approve the request before the agent is provisioned. Approved agents receive a documented purpose and an owner, with access limited to what they need.

The release also adds integrations with MCP gateways, including the open-source agentgateway project. These integrations add identity checks and policies for agents and MCP servers. Supported clients include Claude, Cursor and ChatGPT.

The update supports four authentication methods: private_key_jwt, mutual TLS, SPIFFE and workload identity federation. Internal agents can authenticate through an organization's existing identity provider. Policies are checked on each request, and administrators can narrow or revoke access centrally without reissuing credentials.

The launch comes as AI use on corporate devices expands. In 2025, 45% of employees regularly used AI on those devices, up from 15% in 2024. Among users accessing AI services on corporate devices, 67% used noncorporate accounts.

Descope co-founder and CEO Slavik Markovich said customers had raised concerns about employees “spinning up agents faster than anyone could track them.”