Products & Tools

Virtana puts AI observability agents into Teams, Slack and Claude

Virtana's Teams, Slack and Claude integrations let select customers trace incidents, assess impact and approve governed fixes. General availability is planned for Q1 2027.

Virtana brings AI observability agents to Teams, Slack
Virtana brings AI observability agents to Teams, SlackPexels / pixabay
By Rebecca Stone5 min read

Why it matters

  • Virtana announced Teams, Slack and Claude integrations for select customers on November 20, 2026.
  • The agents trace incidents across five areas: applications, databases, compute, network and storage.
  • Existing identity and access management policies govern visibility, actions and human sign-off.
  • The platform uses more than 25,000 operational metrics and a system dependency graph.
  • General availability is planned for the first quarter of 2027.

By Rebecca Stone

Virtana made its AI observability agents available to select customers in Microsoft Teams, Slack and Claude on November 20, 2026, with general availability planned for the first quarter of 2027.

According to Virtana, the integrations extend its Agentic Observability Enterprise Management product and let engineers and IT leaders ask questions in natural language inside chat threads. They can review affected services and approve a recommended fix without leaving the conversation.

For enterprise IT teams, the immediate benefit is less context switching. Virtana says the launch is intended to reduce the need to move between chat tools, observability software and ticketing systems while an incident is unfolding.

What can Virtana’s agents do inside chat?

Virtana says its agents can trace incidents across applications, databases, compute, network and storage. That coverage lets engineers follow a failure through several layers of infrastructure instead of stopping at the first alert. Virtana says its enterprise policy controls determine which actions require human approval and record actions for audit purposes.

Virtana describes a typical order-processing case in five steps:

  • An engineer asks why the order-processing application has slowed down.
  • The agents trace the issue across systems.
  • They identify a database bottleneck linked to a recent deployment.
  • They show which services are affected.
  • They carry out remediation after approval in the same thread.

Virtana presents the sequence as one chat-driven workflow. The recommended correction stays linked to the evidence and the approval that made it actionable.

How do the agents coordinate and enforce control?

The integrations are built on Virtana’s System-Aware MCP Server and the Agent2Agent protocol. According to the company, those components allow specialized agents from Virtana and third parties to find one another, share findings and coordinate actions across an enterprise environment.

They also use existing enterprise identity and access management systems. Virtana says policies determine what users and agents can view and do, and when human sign-off is required. That arrangement separates discovery from permission: agents can assemble evidence across systems, but policy decides what they may inspect and whether they can carry out a correction.

Paul Appleby, Chief Executive Officer of Virtana, framed the product around collaboration:

“The future of autonomous IT is intelligence available wherever people work. We've built AI agents that understand how complex enterprise systems work and collaborate to solve problems no single agent could resolve alone.

By bringing that intelligence into Teams, Slack and Claude, we're giving enterprises the freedom to work with specialized AI agents through the tools they already trust, accelerating incident resolution while maintaining enterprise control and governance,” Appleby said.

Virtana’s statement links the chat integrations to two enterprise requirements: faster incident resolution and maintained control. The company says specialized agents can collaborate on problems that exceed the scope of any one agent. In Virtana’s design, Teams, Slack and Claude serve as the access point, while identity and policy controls determine what happens after a request enters the system.

Why move incident response into chat?

Virtana positioned the launch as part of a wider effort by software providers to place operational and support functions inside collaboration platforms workers already use. For IT departments, the intended result is a shorter path from a question to a reviewed action. The change also brings AI agents closer to production systems, which makes approval and audit rules central to the product.

Virtana says the same experience will be available across Teams, Slack and Claude. Users can investigate problems and approve corrective actions within those environments. The company expects the integrations to reduce the need for an engineer to leave the conversation while an incident moves from diagnosis to repair.

In Virtana’s model, conversation is the interface for governed work. An access policy can require a person to sign off before an agent remediates a service, and the platform records the action for later review. The control layer remains attached to the request as it crosses applications, infrastructure and cloud resources.

What data sits behind the agents’ answers?

Virtana says its platform draws on more than 25,000 operational metrics and a system dependency graph that maps relationships between applications and infrastructure. The company says this combination allows agents to assess business impact as well as technical root causes during an incident.

The order-processing example depends on those layers of context. Virtana says the agents can connect a recent deployment with a database bottleneck, identify the services affected and present a remediation for approval. In Virtana’s system, the dependency graph gives the investigation a path through the system, while the metric set supplies operational signals.

Amitkumar Rathi, Chief Product Officer of Virtana, tied the chat interface to the quality of the underlying system data:

“A conversational interface is only as useful as the intelligence behind it. When someone asks a question in Teams, Slack or Claude, they're engaging agents that understand how apps, infrastructure, cloud and AI factory fit together. That's the difference between getting an answer and getting a governed resolution,” Rathi said.

Rathi attributes the result to agents that understand how apps, infrastructure, cloud and AI factory fit together. The chat surface provides access to that system knowledge, while IAM and policy controls govern the resulting action.

When will the integrations reach general availability?

Virtana says select customers can use the integrations now, and the company plans general availability in the first quarter of 2027. The release announcement did not name launch customers, pricing, performance benchmarks or the full set of remediation actions available to each customer.

The selected-customer stage gives Virtana a controlled period before broad release. It also leaves enterprise buyers with a concrete date against which to judge whether the same workflow can support diagnosis and approved remediation across all three chat environments.

For Virtana, the next test is whether its governed, chat-based incident workflow can scale from selected deployments to the first-quarter 2027 release without weakening the human approval and audit controls at its center.

Source: itbrief.com.au

Share this article:

More from Rebecca Stone

Rebecca Stone

Show full bio

Correspondent covering companies, deals and strategy in artificial intelligence and machine learning at AI In Context.

802 articles

Related articles

  1. Virtana’s AI agent war room targets cross-system IT incidents
  2. AI Agents Move Onto Windows PCs, and Permissions Become the New Battleground
  3. OpenAI Says Internal AI Monitor Caught Every Employee-Reported Misuse Case
  4. 87% of surveyed tech leaders report AI agent use, oversight lags
  5. OpenAI adds Lockdown Mode and risk labels to ChatGPT

« Previous articleNext article »