Connectors are what make Claude genuinely useful at work — they pull context from the tools your team already uses. The catch so far was the overhead: first an admin enabled the connector for the org, then every single user had to authorize it again themselves. That second step is exactly what Anthropic just removed.
Provision once, available automatically
With the new Enterprise-Managed Authorization, admins set up MCP connectors centrally through their identity provider — starting with Okta. Users then get access automatically on first login, inherited from the IdP groups and roles they already have. The result: zero-touch setup. Open Claude, and your connectors are right there — consistently across Claude chat, Claude Code, and Cowork.
What’s interesting is that this isn’t a proprietary side road. Anthropic built it as an open extension to the Model Context Protocol — the first implementation of the ‘Enterprise-Managed Authorization’ extension. Any connector can support the standard, including the ones your own teams build.
Security that fits the existing stack
For admins, though, the real win is governance. MCP access moves into the same workflow that governs the rest of the stack: provision once, scope by group, manage revocation through the IdP. Because checking access with the identity provider is frictionless, admins can shorten token lifetimes without hurting productivity — when someone is deprovisioned, their connector access expires fast instead of lingering on an old token. Optionally, you can require that a connector only ever connects through the IdP, cleanly separating work and personal use.
A few names are already on board at launch: on the MCP side, Asana, Atlassian, Canva, Figma, Granola, Linear, and Supabase support the standard, with Slack to follow. Early customers like HubSpot, Ramp, and Webflow are already rolling out managed authorization. It’s available now in beta for Team and Enterprise plans.
My take
This sounds unspectacular, but it’s an important maturity step. Connectors are what turn Claude from a chatbot into a work tool — and in large organizations, that’s exactly where onboarding friction kills adoption. The fact that Anthropic built the fix as an open MCP standard rather than a closed feature is smart: it makes the whole ecosystem interoperable instead of just serving Anthropic customers. Identity governance is rarely sexy, but this is precisely where it’s decided whether a tool actually lands in the enterprise.
Sources: Anthropic: Enterprise-managed auth, Releasebot: Claude