
Alex, our IAM Digital Employee, brings Claude Enterprise under the same governance as every other enterprise application. Through the Compliance and Admin APIs, Alex inventories members, custom roles, groups, and the connectors each role grants, then enriches every account with identity context from across the enterprise through Twine’s data fabric.
This makes the blind spots visible: accounts provisioned outside SSO, users whose connector access is out of line with peers in the same role, and applications reachable through Claude connectors that were never onboarded into the customer’s IGA platform. Where remediation is warranted, Alex prepares the change and routes it for human approval within Twine before anything is applied, whether that is removing a departed member, adjusting a role, or updating group membership.
The result: Claude Enterprise becomes a governed application rather than an exception, which is what security teams need before they can expand it across the organization.
Enterprises expect Claude access to be governed by SSO, but accounts get created outside it through pilots, one-off invites, or direct admin provisioning. Those accounts survive termination, leaving active access to corporate data for people who have left. Today Claude sits outside the access review, so nobody is asking the question.
Alex closes this gap because he already governs the surrounding estate: he holds the identity system of record alongside Claude, so he can tell that an account has no SSO path, or that a Claude connector reaches an application IGA never onboarded. That same position unifies access requests, letting a user request Claude access through the company’s existing ITSM workflow (ServiceNow, for example) while Alex resolves the request and provisions the connector, rather than routing it through a separate process. Neither is possible from inside Claude alone.
Step 1. Anthropic Compliance and Admin APIs. Full directory snapshots of organizations, users, roles, groups and settings, plus the incremental Activity Feed.
Step 2. Alex correlates, reviews, remediates. Accounts are enriched through Twine’s data fabric, findings are raised, and every change waits on human approval in Twine.
Step 3. Enterprise IGA and ITSM. Claude appears in access reviews, and access requests run through the existing ITSM workflow, or through Twine.
Nothing is applied without human approval in Twine.



CISOs face two connected pressures 1. Attackers using AI to move faster 2.Internal AI agents gaining access nobody fully governs. Both trace back to the same identity problem, and both need the same fix.
Guilad Regev

Twine Security Co-Founder Justin Woody joins the On Security podcast to explain why AI Digital Employees, not more tools, are now essential in cybersecurity.

Twine CTO Nadav Erez joined Noa Zilberman at AWS Summit Tel Aviv to share why Twine’s IAM customers “won’t go back to the old way”.

Answer twelve questions about what your identity program does today, and the scorecard tells you whether the last mile is yours to build.

How the Twine + Cyera integration closes the loop between data risk discovery and identity remediation.

Twine Co-Founder and CEO Benny Porat joins Jim McDonald to explain how Alex, Twine’s AI Digital Employee for IAM, closes the execution gap.
Connecting the Compliance API brings Claude Enterprise into Twine’s data fabric and lets you govern it like any other application.
Once setup is done, allow up to 24 hours for the data to import. You can then start a new chat with Alex and use the integration to:
Check you are entering the correct API key with no trailing space, and that it is the Compliance API key rather than the Admin API key.
Data from Claude Enterprise is harvested once a day at midnight UTC. Allow up to 24 hours after connecting the integration, and for any change made in Claude Enterprise to appear.