Governed AI assistant integration for enterprise IT operations

Today we're going to demonstrate the MCP server we provide as part of the VOSS product suite. This gives you a standard way to quickly and easily plug VOSS capabilities into any tool that supports MCP. Your admins can tap into VOSS right from the AI tools they already use, and you can just as easily wire VOSS into your own AI workloads or agents as part of a broader automation fabric across your organization. A big benefit of this approach is that it's simple to set up with no integration planning or coding required. I've simply added the VOSS MCP server as a connection and authenticated to VOSS, and that's it. Of course, security and auditability are critical here. Because the connection is tied to authentication, every action is traceable back to the specific admin who performed it, with data segmentation and role based access controls fully enforced throughout. Let's take a look at it in action. The main scenario we're going to show is onboarding a new user, Pauline Harris, through Claude. Let's take a quick look at Pauline's current state in the system. She doesn't have licenses or voice services yet. To change that, let's initiate onboarding in Claude with a simple prompt. You'll see it's reached out to the MCP server and discovered the onboarding tool. This is the power of MCP in action. Claude automatically discovers the tool and consumes its specification without us writing any integration code. It then is looking at what information was provided by the user against what data it needs to finish the request. So you see it's highlighting that it needs a site, and offering some options. I'll provide the site information. Here, we'll use Los Angeles. Next, it notices the user is already in the system, which is correct as we sync it from Entra. So, I'll just tell it to use the existing user. Next, it asks me to provide a subscriber profile needed for the request. We know Pauline needs to be a knowledge worker, so I will provide that input. It doesn't even need to be the exact profile name. Again, the agentic nature of the tool lets it bridge that gap by interacting directly with VOSS. That's it. Now the request is initiated into the VOSS platform. This is a great illustration of the benefits of this approach the orchestrated workflow VOSS already runs can just as easily be triggered by an external AI tool as by a human admin. And if we go back over to the VOSS system, we can see the request was created on behalf of the admin and is being actioned. As this is a full user setup from licensing to voice setup, this will take a few minutes to complete. Now the transaction is complete, we can see the range of steps VOSS orchestrated. And Pauline now has a full suite of services, including voice, ready to go. With that complete, let's look at a couple of other use cases, querying data and interacting with Wingman through the MCP server. For instance, we can ask "how many available numbers do I have in my inventory?". Again, it maps my request to a tool in the MCP server and interprets the response to give me that information. So it remembers the Los Angeles hierarchy that we were just onboarding into and tells me that I have 97 numbers left there. I can also ask for wider context and even have it generate a chart of numbers across all the sites. We can see it pulling the relevant details together from VOSS to build the chart. What you've just seen is the same story we opened with, now in action. With a standard MCP connection, no custom integration work, and every action still fully governed, audited, and tied back to an admin. Whether it's Claude, Co Pilot, or an agent you build yourselves, VOSS is the intelligent automation behind it, finally enabling true business process automation across a range of domains.