Tom Arbuthnot: Welcome back to the Teams Insider Podcast. This week we have Javed Khan, who's CEO of Neat. We get a little bit about his history, his previous role in Cisco, and also his stint in AI and autonomous vehicles. And then we get into what's going on in meeting spaces, his perspective, AI in the room, a little bit on manageability, APIs, and MCPs, and where he thinks that's going, and also a little bit about the Microsoft relationship and what Neat are doing with MDEP. Many thanks to Javed for jumping on the show and sharing so much, and also many thanks to Neat for being such a big supporter of the community. On with the show. Hey, everybody. Welcome back to the show. Really excited for this one. Been talking to Javed for a little while in his new role at Neat, off, off the pod. So this is the first time on the pod. And obviously Javed has a big history in our space as well, which we'll get into. Javed, welcome to the show 

Javed Khan: Thank you for having me, Tom. Looking forward to our conversation today. 

Tom Arbuthnot: Yeah. Awesome. So, for the very few people who don't know who you are, 'cause I expect most people in our community probably do, but, can you give a little bit of, y- your background and, I'd love to h- hear a little bit about this au- autonomous car venture where you came out of Collab and now, now you're back to Collab. 

Javed Khan: Yeah, sure. So, for those of some of you know me, so, you know, I, I, I, I'm new to the new CEO of Neat. I shouldn't say new anymore because it's been four months which is a lifetime, you know, when you're talking about a s- smaller company that's moving fast. But I'm not new to Collab. I spent a lot of time in collaboration, when I was at Cisco. I was running their collaboration business, up until twenty twenty-four. A- and, that was a much larger business, not just the devices business that, Neat is focused on, but the broader collaboration portfolio there. So very, very familiar with the space. We spent a lot of time in the space during COVID, which was a fascinating time for our industry. My background before that was enterprise security software. But I did go off and spend a few years, building a t- and for com- working for a company called Aptiv, which builds autonomous driving software, not just for cars, but for, other, other edge devices like robots and so on. And that's a fascinating space and, and, you know, I think a l- little bit of relevant to, edge devices, and I consider video conferencing devices a form of an edge device, right? But yeah, that space was, probably a, an underappreciated space because when you think about, intelligence running on edge devices versus a data center, cars are probably the most advanced, and robots nowadays, right? Most advanced because they've been doing this for a while. So lots of good learnings there. I enjoyed that space. But when this role opened up at Neat, you know, Neat's taken a very interesting approach with AI at the edge, I couldn't, pass up. So excited to be back in the industry. I think we are going through a second wave of innovation and looking forward to talking more about it today. 

Tom Arbuthnot: Yeah. And is, is that some of... 'Cause obviously AI has happened and impacted lots of spaces during that time, but is that one of the things that brought you back to Collab and into Neat, what, what they're doing in that space? 

Javed Khan: Yeah, absolutely. And, and, and I get asked, you know, you were doing AI for a while on video devices, in the collaboration space. So you've been talking about it, but, something, changed here the last few years, right? And, and, and when I say that, I'm referring to the amount of compute that you can pack into these devices, the kinds of lang- you know, AI models, the smaller AI models that you can pack into these devices. That is a fundamental shift that's happened, in the last technology shift that's happened in the last few years. So we can now do things that we had to do in a data center or, assuming you could even do it in the data center. So I feel we're about to go through another round of innovation because of the technology shifts that have happened from a hardware and from a, from an AI model perspective, which will allow us to deliver a whole new set of experiences in the conference room that were previously just not technically possible, right? And that did-- that was one of the driving factors. I think Neat specifically took a distributed AI approach in the conference room, which is very-- which is different from how the industry has historically approached it. So this coming together of certain technologies, was a big driver and, you know, we've continued to deliver better video meetings experiences, but I think there's so much more we can do in the conference room, and that was a big driver, and I'm, I'm, I'm excited about that. 

Tom Arbuthnot: Yeah, it's, it's, it's super exciting time because the, the models are advancing so fast and not just the giant frontier models, but what you can do with the SLMs, if you like, the smaller language models on the edge, and it's things like, AI direction, intelligence, audio optimization, recognition. I feel like, certainly talking to people at Microsoft, there's a renewed recognition that actually there's a lot of value of doing this at the edge, particularly we get into multi-camera and, and modular systems. 

Javed Khan: Yeah. Yeah, I think there's the advancements in technology, and there's two other things I learned, frankly, from my, experience in the, on the auto and the, and the robotic side of the house, when I was at Aptiv was you also have a latency problem that's getting solved, all right? Because it's one thing to be able to run these things locally. You, you don't have the network delay that you would have had if you were running them in the data center. And the third benefit is privacy, right? You can now keep data, keep, private content local. It doesn't leave your campus, right? So those are two other problems, if I may, that now this technology shift is allowing us to solve for. And, I think, that, that will allow us to do certain things that were previously not possible, right? I mean, even the simple things around sometimes customers were nervous about transcription and translation in real time, even though we've been doing that in the cloud for a while. You now have the opportunity to eliminate the latency problem that, cloud solutions have and the privacy problem. We can now confidently say it's all running locally. 

Tom Arbuthnot: Nice. So taking it a bit wider, what's your current kind of thesis or thinking around meeting spaces? 

Javed Khan: Yeah. So I'll start with, with, a Neat view of, of the world, right? So, you know, we are, no longer a startup, you know, starting to grow pretty fast. So, when I just think about where Neat needs to go, we, we are focused on growing our business, right? So at the end of the day, we, we made a lot of progress with small, medium rooms in our early, stages. We added, you know, started off as a Zoom shop and then expanded to other platforms. So focusing on growing the business. So that's, that's one of the core focus areas for Neat as a company. We've, we're seeing, a lot of hypergrowth, but we wanna make sure we are, building a company, that is, ready for the scale that our customers expect from us, that the market expects from us. I think the second, meeting space, ask I'm seeing from customers is, they are, spending their time-- They made their platform decisions, right? Whether they use Zoom, Microsoft, Google, they are looking for better experiences, as they go through their day w-when they are asked to move between platforms. So they love the experience they get from, from, from, from these devices. They obviously want those to be, better, leveraging AI, but, make it easier for me to go, go through my day when I join another meetings platform. So working with the, various, platform partners, right? Microsoft Teams, Zoom, to make sure that that experience, gets better. And then the third thing is they continue to want simplicity with managing this ecosystem of theirs, right? Because they have multiple vendors oftentimes, right? They are using it not just for, for, for video conferencing, but increasingly the devices are being used for, for other internal use cases, right? Wayfinding and, internal messaging and so on. So they want to simplify the management of that ecosystem, and, increasingly do that using some of the agents that they're starting to install rather than having to put humans, every time. So that's, those are some of the challenges, that I think the industry has to continue to work on. I would also add the underlying foundation, which is we cannot do enough to make the meetings experience better, right? That's one of the things that I learned, during my days at Cisco, is let's-- we need to-- there's more we can do to make that experience better. There's still people when they walk into a room, right? If they're not familiar with the ecosystem, there's an extra click or two, there's an extra phone call or two sometimes that's involved. And, and the industry can do more there, across those platforms. 

Tom Arbuthnot: Now that, that's another area where AI is really exciting, like the potential of it being much more assistive. Like conceptually it can understand there are people in the room and actually you've got good voice modeling now, so actually rather than having to press buttons, maybe I can ask questions and get intelligent answers, that kind of thing. 

Javed Khan: Yeah. I think the interface can be a more human interface. I think it can also be a learning interface. So, if you notice that the particular room has the same set of problems every time, whether it's, technology, adoption or it's just there are people sit in the room, or, the, you know, the fact that a, that a, that, that a person, always ends up in the front of the room and is out of camera, right? Or out of sight of the camera. So I think how can we adapt based on problems? I think AI has the ability to learn those things and just make that experience better every time. And, and, and really these are things that previously computationally were just not possible. Right? And you can now start to do these things. 

Tom Arbuthnot: And you mentioned management. That's something I hear a lot from enterprises. As you say, typically there's more than one device partner in play. They've obviously typically got their, their vendor's chosen management platform, Zoom, Google, like Teams. Neat are the first to have a public preview of MCP in, in the space. Mm-hmm. And I, I'm really excited. I'm, I'm spending a lot of time with AI and agents as a techie probably is at the moment, but like the, the potential of doing things through MCP from, from chat or from agents is really exciting. 

Javed Khan: Yeah, I think it's an amazing piece of the, amazing development, frankly, in this agentic AI space, which, not only is going to make the interface into these management systems easier because it's more of a human language interface. And we, you know, we built an MCP server where we are exposing the data that we have, in our management, systems to these agents via MCP. You can now interface using a more human-friendly fashion. But I think the other thing that this technology is going to allow us to do is bridge disparate systems, whether it's disparate management systems in the video conferencing space or other kinds of enterprise applications. So I can now very easily connect what comes out of, you know, whether it's problems or data coming out of, the management console, Neat Pulse, that we have, with ServiceNow, or with Salesforce, or with another ticketing system you might have. And I don't have to worry about a direct integration with them as long as I support, a, a MCP type protocol. So what we are doing is we are saying, okay, you know, yes, we've had Neat Pulse, customers love it, but it still involves you going in there, and, and, and selecting certain bells and whistles. You can now do that by having a conversation, right? And then hooking that up, to the agents. As an example, you wake up in the morning and, you know, any systems that are not quite working, you can now automatically open up a ticket and that, that kicks off its own workflow. Um- 

Tom Arbuthnot: Yeah, that's what I want, is the agent doing the first pass for me rather than me having to go into Pulse or any other tool in the morning and, and scan down and check or get email alerts. Why is the agent not doing that? And then potentially, you know, we're there where like, actually we know what the typical remediation is for that. We're gonna do A or B or C. Like if you've done it three times, let's just get the agents do it 'cause that's what I'm gonna do anyway. 

Javed Khan: Absolutely. Yeah. So I think the-- So that's why we are also looking at making it a two-way, process. So it's not just getting data out of our systems, but, but, but taking action, right? Yeah. Because that's where the true power comes in. You obviously have to protect it from a security and, have the access control in place. But that's how we are approaching it. You know, you can send a message out to the device and say, "I'm about to take certain action, or, or reboot the device," or so on. So that's how we've approached it. We've, we, we look at ourselves as, you know, we, we are really, really good and focused on what goes on in the conference room, not just the experience, but also the manageability of it. Let's take that data, that learnings, that intelligence that we are gathering from the conference room, whether it's while you're in a meeting, before or after, and make that available, to these agents. And what you do with that, others are really very good at that, right? So let the, agents deal with the language interface. Let's, let's them deal with the connection to other enterprise systems rather than us trying to build all of that out, which is what- Yeah... some of us tried to do earlier. The other thing I'd add is, it also allows us to hide a lot of the complexity around management. So historically, we-- companies have had to build, lots of bells and whistles. You know, what is the, latency of a particular s- what's the jitter is a particular subcomponent performing? You can now not worry about exposing that directly, because this interface, abstracts that from you and, fix the system versus telling me about what the latency is. So we don't have to worry about all those bells and whistles, just significantly simplifying the c- the, the ability to manage these devices. 

Tom Arbuthnot: Yeah. I mean, we, we, we know it's so hard to build an interface for these kind of things because every customer, you know, someone wants every piece of latency metadata on the screen, the other person wants none of it, and they want simplicity. So the ability to have flexibility there to be like, "I can go in and get the data if I need it, but actually I can have a very simple interface if I need that." And even, I don't know, we're doing a lot of building, you know, custom reports with AI from various different tools and, like, the ability to generate the exact UI you want is really exciting. 

Javed Khan: Yeah. I see, and this is me getting a little philosophical, right? But as I see AI sometimes, y- you know, you can use AI to replace existing complexity. But the bigger, power is to eliminate this comp- completely, right? Which is we-- I, I needed these bells and whistles in the past because of a certain reason. But if I can have a higher language, higher level conversation and s- and, and s- and, and eliminate it completely, right? Why not do that? And that's the approach we've taken, which is use technology not j- not to replace an existing complex piece with a slightly simpler human interface, but can you imp- can you eliminate that completely, whether it's manageability, joining a meeting, or the experience element? 

Tom Arbuthnot: Nice. And talk me through a bit about your perspective and Neat's perspective on multi-platform, 'cause Neat, you know, came up on Zoom initially, made a massive push into the Teams space and now is a, you know, big part of the business I believe is now Teams but Google's now on the scene as well and BYOD modes. 

Javed Khan: Yeah. So, we, are, you know, in some ways what I call is the Switzerland of the, of the, of the d- devices industry, right? We don't have a meetings horse in the race or meetings platform horse in the race. And by the way, they're more than meetings platforms nowadays, right? So these are collaboration platforms. They do so much more than just video meetings. So we, so, so we, we don't, you know, we are not making a bet on, on, on those platforms, but we want to support every platform that our customers are using. So we did start off with Zoom. Obviously coming out of COVID, there was significant adoption of Zoom, but there was also Microsoft Teams adoption. So those were the two platforms we supported. We are increasingly seeing our customers ask for, Google, so we'll continue to support that platform. The in-meeting experience that we run is 100% certified, supported, enabled by software from those vendors. We, we, we, we don't try and, you know, influence what happens when you're in that meeting. And, And we will continue to support, newer platforms, newer platforms as our customers ask for it. What the value we provide is, the experience, the video experience, the audio experience, the, the quality of that experience, from the room, making that better. So, our camera technology, our audio technology, the ability to, deliver that better experience, is one piece. The second is, a big part of the cost is often setting up these ecosystems, setting up these rooms. And so we spend a lot of energy on the process of setting up your room. You know, we- everything's over IP. It's a distributed AI architecture. You can manage that through Pulse, now MCP. So that's the other big value add because oftentimes it's the cost. You know, you can order one of these devices, have it show up, but what does it take to get it up and running? You know, do you have to spend an extra $30,000 for a $10,000 you know- 

Tom Arbuthnot: Yeah, yeah, yeah... hardware spend? And the kit device, the, the, the devices have got much more cost efficient over time, but the install cost can often outweigh that if, if it's a complex install or, you know, needs special configuration and special programming, all that kind of stuff. 

Javed Khan: Absolutely. Yeah, so that's, so that's the value add we provide. But, but we're gonna stay true to we don't have a meetings platform in the race unlike some of our competitors. It, it just, makes it so that the base level functionality we build forces us to make that work across all platforms, versus being, you know, having certain things on one platform versus not on others, so. 

Tom Arbuthnot: Makes sense. Makes sense. And, like I, I guess coming back into the space, you've, obviously picked up the relationship with Microsoft. You've got your first MDEP device out there. So, we're obviously a Microsoft-focused community. It'd be great to get your take on Microsoft and the Microsoft relationship. 

Javed Khan: Yeah. So, so, we've been partnering with Microsoft for a while. Again, Microsoft's done a great job of, you know, bringing, frankly the early stages while there was a lot of Zoom adoption, the wider enterprise adoption has been driven by the great work Microsoft's done with Teams. Platform continues to get better, so we continue to work with them closely on enhancements they are making, not just with things like MDEP, but you know, this, facilitator that, that's, that's increasingly, being talked about. With MDEP, we are starting to see, the manageability get a lot more easier in the Microsoft system. So as Microsoft pushes that across the ecosystem, not just within video conferencing, but other, devices, I think, the concern that customers have had about what version of Android or what OS am I running on this hardware- Mm-hmm... how do I manage that, how do I update it, will get significantly better. So we- we've, we've signed up to support that across our portfolio. As we ship newer devices, we'll enable that. I think it's going to be, if you're a Microsoft shop, right, it's gonna be a great way. It's gonna simplify, your s- your, your manageability and security. I think security is the other big driver. So we are excited to be a part of that ecosystem. It will also allow us, over time as the platform gets adopted, to make our development a lot easier because now there's an underlying base OS, that's, that's standardized across, our, our, our hardware, so our functionality make it work once and it'll work across the Microsoft ecosystem. 

Tom Arbuthnot: Awesome. Well, Javed, thanks so much for jumping on, and just on a personal note, thanks for, you know, you spend a lot of time with the analysts and the community, and you're always willing to answer questions and stuff, so I appreciate the, the level of engagement you have with us. 

Javed Khan: Yeah. Well, thank you. Thank you so much. It was a great conversation and, looking forward to more of those. Yeah. 

Tom Arbuthnot: Yeah, yeah. We'll give it, we'll give it another, we'll give you another six months or maybe, maybe your first year in and then we'll revisit if you like. 

Javed Khan: Absolutely, yeah. 

Tom Arbuthnot: Awesome. Thanks so much, Javed. 

Javed Khan: Thank you.