LTM AgenTraceIQ Lists Copilot Studio Support, but Admin Integration Details Are Missing

What LTM announced​

On October 5, LTM announced BlueVerse™ AgenTraceIQ, an offering meant to help organizations securely adopt and scale agentic AI. It combines Rubrik Agent Cloud with LTM's AI governance and managed services expertise. The stated goals are to monitor AI agents, set guardrails, and rewind unintended agent actions.

LTM's product page splits the offering into three capabilities:

• Agent Monitor: discover agents, see how they interact with data and applications, and keep an audit trail.
• Agent Govern: apply policy-based controls over agent behavior and tool or data access, using zero-trust principles.
• Agent Rewind: restore systems, data, configurations, and repositories to a defined state after destructive or unintended actions.

LTM also says it can trace actions back to prompts, plans, and tool calls to help with root-cause analysis. The service layer has four parts: AI Resilience Assessments, Governance Framework Design, Deployment & Integration, and Managed Governance Operations.

Where Copilot Studio fits​

The offering says it supports Microsoft Copilot Studio, Salesforce Agentforce, Amazon Bedrock, Anthropic Claude, and custom-built frameworks. That is the Microsoft hook. The announcement doesn't say which of the three capabilities work with Copilot Studio, or whether they work the same way on every listed platform. It gives no supported versions, prerequisites, licensing terms, or regions.

The Rubrik and Project Hourglass background​

This launch grows out of an alliance Rubrik announced on June 9, 2026. Under Project Hourglass, LTM and other Global Systems Integrators deliver Rubrik Agent Cloud for Anthropic's Claude Code to enterprise clients. Rubrik's own announcement named Cognizant, Deloitte, LTM, HCLTech, NTT Data, and Wipro as launch partners. It said the program gives integrators joint go-to-market support, technical certification tracks, and access to Rubrik engineering teams.

Rubrik's detailed technical claims concern Claude. Its Claude Code and Claude Cowork release describes several features:

• Agent inventory.
• Agent Rewind.
• Codebase resilience: snapshots of GitHub and Azure DevOps repositories, stored outside the repository and beyond the reach of compromised credentials. Rubrik says this covers cases such as force-pushing over commit history or deleting every branch.
• Configuration backup: it backs up the configuration that governs how Claude agents behave, including system prompts, tool permissions, skills, and files such as CLAUDE.md.

Nothing in the LTM material says these Claude-specific mechanisms carry over unchanged to Copilot Studio, Agentforce, or Bedrock. Treat the Hourglass description as background on the Rubrik engine. It isn't a feature list for every platform LTM names.

How it compares with Microsoft's own controls​

Copilot Studio already has a substantial governance stack, and an outside service would sit next to it. Microsoft Learn's security documentation describes several layers:

• Data policies. Admins use the Power Platform admin center to govern maker and user authentication, knowledge sources, connectors and skills, HTTP requests, channel publishing, and triggers.
• Auditing. Maker audit logs are available in Microsoft Purview. Admins can also monitor and alert on agent activity through Microsoft Sentinel.
• Agent 365. Microsoft describes it as a central control plane. For organizations that onboard it, Copilot Studio agents can be represented as Entra identities and governed with Conditional Access and access governance workflows.
• Source control. GitHub-backed source control and deploy-from-Git provide an auditable deployment history.

The Learn page also carries a caveat. Copilot Studio security audit telemetry runs through the Purview audit pipeline rather than the Copilot Studio service. Certain Agent 365 governance and audit events flow through that separate control plane. Customer Lockbox doesn't cover either category. So an admin adding a third-party layer is dealing with at least three telemetry paths.

None of this shows that AgenTraceIQ replaces, duplicates, or integrates with any Microsoft control. The sources don't establish any of those. This is a comparison of scope, not a claim of equivalence.

What the announcement doesn't say​

The release and the product page leave out:

• Pricing, availability dates, and onboarding timelines.
• Customer names or case studies.
• Service-level commitments.
• Any independent testing.
• What "continuous" visibility means in collection frequency or latency.
• Whether recovery is automatic or needs human approval.

LTM's page says recovery can happen "in minutes," but I found no test data or defined conditions behind that figure. It is a vendor claim. Also note that the copy is promotional. The statements from Rubrik and LTM executives are positioning, not evidence of results.

Questions to ask before buying​

Based on the stated scope, a Microsoft-centric team should get written answers to these:

1. Is our Copilot Studio agent inventory covered in our tenant and environment setup, including agents in separate Power Platform environments?
2. Which prompts, plans, tool calls, and audit records are captured, and where are they stored?
3. Which actions can be blocked in real time, and which can only be reversed afterward?
4. What can be restored, to what point, and who approves a rewind?
5. How does this interact with existing Power Platform data policies, Purview auditing, and any Agent 365 deployment? Which one wins when they conflict?
6. Who owns incident response when an agent misbehaves: the integrator's managed service or your own team?

Bottom line​

The concrete news is a packaged service. LTM is bundling Rubrik Agent Cloud with advisory, integration, and managed governance work, and it says Copilot Studio is covered. Rubrik's Claude-focused material is more detailed than anything published about Copilot Studio support. Until LTM documents the Copilot Studio integration and shows it working, ask for a proof of concept in your own environment.