·
4 minute read
Summary
On this New Product Alert, the Collab Collective’s Craig Durr speaks with Kevin McMenany, Chief Architect – AI Integrations at Cisco. Recorded at WebexOne 2026, the conversation explores Cisco’s Collaboration AI PODs, an on-premises AI solution designed to bring Webex collaboration capabilities to organizations with strict security, privacy, and data residency requirements.
Their discussion covers:
- On-Premises AI for Collaboration: How Cisco adapts its AI POD infrastructure to deliver Webex collaboration AI capabilities within customers’ own secure environments
- Core AI Capabilities: An overview of automatic speech recognition, meeting summarization, and text-to-speech functionality for meetings, calling, and contact centre applications
- Data Security and Control: How on-premises AI inferencing gives organizations greater control over sensitive meeting data, processing locations, and recording storage
- Secure Webex Architecture: How Video Mesh integrates with end-to-end encrypted meetings to enable transcription, recording, and summarization within trusted enterprise infrastructure
- Future AI Developments: Potential on-premises expansion into AI-powered calling, contact centre IVR, AI Receptionist, real-time translation, and enterprise knowledge capabilities
Transcript
Craig Durr: Hi everybody, it’s Craig Durr, Chief Analyst of the Collab Collective, at WebexOne 2026. Here is more great, interesting stuff coming from the expo floor. This is one we talked about a little bit last year, but there have been a lot of great developments since. They're calling it AI PODs, and that's the hardware you're looking at. What's more important is the feature set they're bringing to this solution. So who is the best person to walk me through this? I think it should be the chief architect, so I'm going to ask for Kevin McMenany.
Kevin McMenany: Good to see you again, Craig.
Craig Durr: Kevin McMenany
Kevin McMenany: You got it.
Craig Durr: I just like saying Chief Architect. It just works well.
Kevin McMenany: It's easier.
Craig Durr: Kevin, you and I have known each other for a while, so I appreciate you letting me joke. I'm so excited about this. You start with the hardware solution. Why don't you tell me what we have here?
Kevin McMenany: This is a Collaboration AI POD. Cisco makes AI PODs in a bunch of different form factors through the NVIDIA partnership that we have. We've taken one of those servers and made it specific for collaboration use cases, and that's really about the software that we put on there. This is basically the collaboration AI features that we have in Webex, brought on premises.
Craig Durr: So this means certain customers, such as high security, military, and sovereign states, can now deploy this on premises with some of the AI features we're talking about here.
Kevin McMenany: Yeah, and we're starting off with three features so far. First is automatic speech recognition, which is the speech-to-text engine.
Craig Durr: Which is a core element of getting us all together.
Kevin McMenany: Exactly. It drives the closed captions you see on the screen, and it drives the transcript you see at the end of the meeting. Next is summarization. Now that we have a transcript of the meeting, we summarize it: who said what, what the action items were, and so on. And next, we're working on text to speech for contact center IVR use cases. So it's powering meetings, calling, contact center, and other use cases.
Craig Durr: And what you're saying is that all the data stays here. It never leaves, right?
Kevin McMenany: Well, all the inferencing is done here. The storage of the recording, for example, is up to the application that's using it. This is just the AI inferencing engine.
Craig Durr: Got it.
Kevin McMenany: The recording would be stored on the Cisco Meeting Server, for example, or the Video Mesh server.
Craig Durr: That makes sense. So in a case where I want to ensure that my meetings and my transcripts are not going off premises, they're not potentially exposed to compromise. There's no cloud exposure taking place.
Kevin McMenany: Right. You know where the data is being inferenced. You know where it's being stored. It's under your lock and key in your secure data center.
Craig Durr: That's great. You actually have an architecture slide here. Come on, let's take a look at this. So we're looking at what this might look like on a customer's premises, right?
Kevin McMenany: This is a Webex meeting, an ultra secure, end-to-end encrypted Webex meeting. But with the addition of Video Mesh, as an administrator of your enterprise, you can say, “I trust this Video Mesh server. I installed it. I know where the keys are.”
Craig Durr: Exactly, it's locked down.
Kevin McMenany: You're going to enable it to participate in this end-to-end encryption, thereby decrypting the media here. Now that it's decrypted, you can record it, and you can transcribe it and summarize it.
Craig Durr: Using our favorite Collab AI POD.
Kevin McMenany: Exactly.
Craig Durr: And then you get the feature set.
Kevin McMenany: Now, you have the best of both worlds: ultra secure, end-to-end encrypted in the cloud, while still having the feature set you're used to on Webex.
Craig Durr: This opens up a whole new world of possible features coming to this audience. What are you thinking about?
Kevin McMenany: In the future?
Craig Durr: Yeah.
Kevin McMenany: Oh my gosh, it's a long list, and we're really excited. All the features we have in the cloud today, like being able to ask it any question. The corporate brain, the cognitive layer that we talked about in all hands, running on prem. That's the big picture thinking. Longer term, we'll get there.
Craig Durr: And you're even thinking about some contact center IVR scenarios as well.
Kevin McMenany: Contact center IVR scenarios, Calling, all of the AI assist features that we have in Calling and Contact Center, AI Receptionist, and the real-time translation. We acquired EzDubs in December.
Craig Durr: That's right, so all of that is possible to run on here. It's just a matter of when it comes together. So the three features you talked about, are those available right now to customers?
Kevin McMenany: Yes
Craig Durr: I love it. This is a great story: AI PODs Collab Edition, a secure architecture, and a roadmap that is very fruitful going forward. I appreciate it, Kevin. This is a great update.
All right, I'm going to sign off. Everyone, this is Craig Durr, Chief Analyst of the Collab Collective. We are here at WebexOne 2026, and we'll bring you more updates shortly. Take care.