AI as the New Operating System: Transforming SMBs, Security Risks, and Unified Experience Platforms with Anurag Agrawal

[00:00:13] Today, we'll explore the idea that artificial intelligence is no longer just a tool, but is rapidly becoming the operating system and foundational platform for modern businesses. We'll cut through the hype to examine what's real, and what's not, about AI's impact, especially as it becomes both a powerful security tool and a new attack surface. Welcoming back special guest Anurag Agrawal of TechIsle. Welcome to the Business of Tech Lounge, the live version of the Business of Tech Podcast.

[00:00:41] It's Wednesday, July 9th, 2025, and I'm Dave Solberg. Welcoming questions and comments throughout the show, so please make sure to submit them in chat. If you have a question, we will happily respond to them. We'll do so over the course of the conversation. Now, I want to thank Sales Builder, our Patreon sponsor whose support makes this show possible. Focus on your IT sales workflow with the power of automation, and visit them at salesbuilder.com. That's B-U-I-L-D-R dot com.

[00:01:10] Reminder, I am watching that chat. Keep taking your questions over the course of the call. Anurag Agrawal is the founder and chief global analyst at TechIsle, a research firm focused on SMB, mid-market, and channel partner IT trends. With prior roles at IDC and Gartner, he's recognized for guiding MSPs and vendors through AI integration, digital transformation, and channel evolution. He's a good friend to the podcast. Anurag, welcome back to the show. Anurag Agrawal, MD, I'm honored to be here, Dave.

[00:01:40] Oh, I'm always excited to have you on. You come with such cool, interesting ideas, and we have such fun discussions. I got to start with one recently. You talked in your research about the idea of AI as the new operating system for SMBs and mid-market firms. Give me a little bit of sense of how you're defining this concept and what it looks like in practical terms.

[00:02:04] Sure. You do keep a tab of the various research papers that we really put out, and you pick up almost the most important topics which others ignore. But yes, the AI as an operating system, what I think is that it is really truly transformative, right? But yet nascent concept, right?

[00:02:28] An AI OS or AI as an operating system would really embed AI at the very core of computing, really managing resources, interacting with hardware and orchestrating applications with really intelligent adaptive capabilities. And I just mean that it wouldn't just run on an OS. It would be the OS.

[00:02:53] That means it would learn user behavior, anticipate needs, automate tasks, proactively blurring the line between user intent and system action. But the challenge here, Dave, is the fact that, you know, this AI as an operating system often gets conflated with advanced AI user interfaces or highly integrated AI applications running on traditional operating system.

[00:03:21] This really isn't true AI OS. Now, I want to get a little bit of your thinking on this because, I mean, your word nanosent like is doing a lot of heavy lifting right there, right? Because I think you're talking about this conceptually, not as something that we're definitely not there now. I mean, I would make a pretty clear argument that even agentic is not really there than its trajectory. Right.

[00:03:44] Your thinking is the idea of we're going to embed actual intelligence into the operating system that you use. And we're trying to use it abstractly, right? I'm not necessarily saying Windows or Mac or I like we're just sort of saying in the systems you use, they should become intelligent to present information at the same time. Do you have like a specific example of what you think that would look like?

[00:04:07] You know, that is really challenging because what I see is that the transition is really evolutionary, right? It will really appear when we will have AI native kernel features and intelligent resource management come to traditional operating systems. Right.

[00:04:30] So the way I see it is at this point in time, the AI OS is really an aspirational destination, not the current state. What this actually means is, you know, giving you an example is, hey, if I start a PC, right? Hey, there is a Windows operating system that starts first, right? Or you have a Mac OS that starts first. And then you have got these platform capabilities, right?

[00:04:57] Which we are still very much in the AI as a platform area, era, right? Where Microsoft with Copilot Plus or Apple with Apple Intelligence, since I picked up those two names here, are integrating AI into their OS as deep as possible, right? There are still enhancements to the OS that needs to be made.

[00:05:18] But then if you look at an AI operating system as an era, then when I really boot the system, it will know exactly what I have to do, when I have to do at the time that I have to do and would know exactly my context, my personality, how at what angle I'm keeping my PC. Is it me or is it somebody else? What is the first action I need to take?

[00:05:46] Can it take that action on my behalf and cut down the time that I need to spend on looking for files or launching into a Zoom call or a WebEx call and things like that? So those are kind of simple enhancements, but then you kind of start to extend it and the possibilities become really limitless. Now, I'm going to try and make this a little bit tangible and I'm going to pull from the environment that always is the most fun and that's always Star Trek, right?

[00:06:14] So if we think about like the Star Trek computer, the idea of talking to the computer and it just simply anticipates things. The best part about that is it blends into the background entirely. You don't think of it as an AI or an intelligence. You're simply interacting with a computer. Does it kind of in the end state that you're thinking about? Is that a good way of thinking about it? Yes, yes, exactly.

[00:06:37] So to put a spin on it, this means that we will start to prepare for an intent driven computing where users like me or you can describe the outcome and the AI operating system orchestrates the how. Right, right. Right.

[00:07:01] And I think that's where it starts to become really, really important and then it will be be very, very fruitful or really, very productive for the partners as well, because then it will involve new consulting services around AI OS adoption strategies, AI native application migration and managing AI complex workloads at the infrastructure level.

[00:07:26] Okay, now I'm going to try and pull it back to something a little bit more close to where we're talking about, right, because I get it as a direction if we think of it that way. But you've also talked in your research about the idea of agentic AI, right, the idea of the evolution from traditional robotic process automation or RPA to more judgment driven automation. One of the things that I'll observe is the real power right now is that the struggle of RPA traditionally is it's just hard to do. It's really hard to do.

[00:07:54] And the wrapping generative AI around it makes it a lot easier. Where do you see from the research you're doing around agentic AI? Like how real is that conversation right now? So our research, at least within the SMB mid market and the upper mid market, is showing that the current adoption may be nascent.

[00:08:21] Exploration is high, you know, starting with automated customer service, IT operations, supply chain, automation, as leading target functions, right? And the reason why this is so is because the way these end customers are thinking about it is that these agentic AI solutions are capable of autonomous multi-step tasks.

[00:08:51] Right. Which is very different than RPA. RPA was, I do not even know why RPA came into the conversation for the last two to three years. Because, hey, it is something that you are automating a manual task. And if anything changes, whether the software changes, application changes, or the user interface of the box or the pixels on the screen change, the RPA starts to fail. Whereas agentic AI is not supposed to fail, right?

[00:09:18] And the reason why this is also coming up in many, many different conversations is because the end customer is looking at the expected benefits. And one of the key expected benefits is about operational autonomy. It's not about just operational excellence, right? And handling complex processes without human intervention.

[00:09:41] And this is important because when you look at it, the maximum amount of time that an SMB customer or a mid-market customer spends is about automation of workflow. And the key area where the SMB customer is trying to look for that elusive holy grail is that automation of that workflow. And agentic AI comes with that promise.

[00:10:09] Now, comes with that promise again is the key word because I have been watching a lot of the work with model context protocol or MCP. Right? And I am sure you are looking at the same thing. This feels both fast and slow at the same time. And what I mean by that is fast adoption. It was surprising to me how fast MCP was adopted as a standard and Microsoft, Anthropic, OpenAI, Amazon, a bunch of people all moving very quickly.

[00:10:36] I mean, there's even a WordPress version of it already rolled out for those that are using WordPress. But at the same time, it's very just rolling out quickly as a standard. But a lot of people are still figuring out how to use it. Is that what you mean when you say like still Nanosent? Like is it that we're still trying to figure out what we can do with it? Yes and no. Right? I mean, yes.

[00:11:00] MCP from Anthropic and A2A from Google have really caught the imagination because they kind of fill a gap. Right? But the reason why, you know, people are kind of still experimenting and they are unsure as to how to adopt agentic AI because the field of AI has moved extremely fast for them. Right? When you had the chat GPT moment from there, you moved on to Gen AI. And now you're talking about agentic AI.

[00:11:29] And there are not many organizations or consulting firms that are really conversant with the end to end technology adoption. Right? I mean, yes, agentic AI can solve the workflow automation problem. But then how do you secure it? Right? What types of agents are these? The co-pilot agents? Are these the bots? Are there something else? How do you secure them end to end? Is it at the context level?

[00:11:58] So I think there are many, many different areas which are still being really understood. You know, for example, our research is showing that defining clear accountability and liability for agent driven decisions or errors is the top challenge. For the SMBs at 77%.

[00:12:24] Those who are actually aware about agent AI are actually moving towards adoption for upper mid market. It is 94%. So who is able to solve that problem today? I don't think there is any organization who's able to solve that problem. I think you're 100% right. And I think accountability is going to end up being the bid. In fact, I had Brad Gross on the live show recently who literally encircled this as like the biggest area of risk was around accountability.

[00:12:53] So give me a little bit of a sense of, you know, you've been you're watching all of this and you're thinking about the various misconceptions and sources of hype. Like what are the most common ones you're seeing about AI adoption in the SMB and mid market space? So see, AI adoption is at the very heart of new technology investments where the SMBs are really investing in. Right.

[00:13:20] And the key area is that, you know, there is pervasive adoption and investment. There are strategic imperatives that are driving AI adoption, which is operational efficiency, automation, automating processes, augmenting employee capabilities, data analysis, insights, and things like that. Right.

[00:13:43] So, but from a vendor perspective, despite the enthusiasm, what we are seeing is the organizations are considerably facing hurdles, right? The shortage of skilled AI talent, the complexity of integrating disparate AI tools, data related issues, like insufficient quality data silos, privacy, security concerns. Right. These are challenges which are extremely difficult to overcome.

[00:14:13] And unless there is a consulting firm that says, you know what, I'm really focusing on the SMB market and I will go and do a readiness assessment and then combine that with hybrid AI adoption. Right. So I think that AI is, is one thing, but looking at hybrid AI is another important area where there is tremendous opportunity for your listeners, your partners and your technology vendors.

[00:14:38] Like because you have to combine different AI techniques or deployment environments on prem cloud and edge. Right. Mm hmm. So these are all becoming extremely important. Those AI readiness assessments are a continuing theme that I keep hearing in all of the conversations that I'm having. And I want to continue to highlight that for listeners. And if you're watching us live, throw a question in the chat. We'll happily take those over the course of the conversation.

[00:15:04] Now, I would be remiss if I didn't talk also about security, which is a major theme in a lot of the research you've been doing. Are there particular generative AI powered threats that are changing the landscape right now? Like what's the most urgent risks that you're concerned about? Well, so, so the way I see it is AI is both a sword and a shield. Right. Okay. It's a sword because the attack surface is there. Right.

[00:15:33] The new AI attack vectors enemy, I can talk at nauseam about this, whether it's, it's about prompt injection or, you know, model poisoning and things like that. But, you know, the, the very important things that are really appearing within the radar, right?

[00:15:55] It's about feeding poison data is the enhanced social engineering, right? Because Gen AI allows these attackers to really craft highly personalized and convincing phishing emails, deep face audio video for voice impersonation and things like that. Right. Right. Then there's automated exploitation, which is there, right?

[00:16:24] Which is AI can automate vulnerability scanning and exploit development, accelerating the pace and scale of attacks. Then there is data poisoning, which is going to talk about, right? Attackers can introduce malicious data into AI training sets, leading the AI to make incorrect decisions or creating backdoors for these attackers. And obviously there is this whole area of supply chain risk, right? AI models often rely on open source components.

[00:16:54] You go to, um, hugging face, right? Which is like the entire repository of LLMs. You know how many LLMs are there? There are over 2 million LLMs on the hugging face. There are well-known ones that are smaller ones, right? Who knows which ones are really, you know, vulnerable and which are not.

[00:17:16] So when AI models often rely on open source components or third party data, introducing potential vulnerability to the supply chain becomes really easy. And those social engineering attacks are not at all theoretical. I mean, literally the Washington Post covered just today. So Secretary of State Marco Rubio being impersonated with active voice calls into into foreign officials and US government officials like this.

[00:17:44] These are legitimate attacks that are happening right now with voice cloning. Uh, I had an expert on the show literally talked about how easy this was to do a year and a half ago. Uh, you know, and I know like for any of us that speak at any length online, it's a real threat, but think about how easy it is to do for, you know, a CEO or an IT manager or a CFO to be impersonated. It's incredibly difficult, but you talked about it being both the sword and the shield.

[00:18:13] Are there particular areas that you're seeing the shield starting to come into play, you know, particular solutions or, you know, better, you know, better approaches that are making a difference here in terms of their use. Dave, you know, as an analyst, my head starts to explode when I have these briefings with the different vendors, because everybody, you know, says that theirs is the best. But I, what I feel is that there are four areas, right?

[00:18:42] One is you have to have this protective threat detection capabilities, right? Where you can use AI itself to analyze massive data sets, network traffic logs, endpoint activities and so on and so forth, then automated response, right? How can AI really, how can you use AI automatically to do the remediation task?

[00:19:07] Isolate compromise systems or then you have got a vulnerability management, so on and so forth. But I think the way you have to think about is going one, one year down the line is that we're going to be able to do this. When the application itself, the software application being used by the end customer is changing, right? It's not just the SaaS applications.

[00:19:27] When you build in this agent capabilities and agent AI capabilities and all that, then it's not only about the one database where the agent is reading and writing information from. It is not just integration between applications. It is a lot more than that, right? So that's where you have to start to think about it. How do you protect that entire area?

[00:19:49] And one of the latest ones, which I've heard is, sorry to bring this up in your podcast, but Palo Alto Networks and they call it the Prisma Airs, AIRS, which is kind of a cloud-based as well as somewhat on-prem solution as well, which kind of looks at all of these vulnerabilities.

[00:20:13] Which, you know, with red teaming and many of these activities, but until and unless the security vendors move away from this point solutions to a platform-based approach. I think Cisco is taking a pretty good step forward in this area as well. I think these two organizations, Palo Alto Networks and Cisco are doing some really groundbreaking work, but they are still at an enterprise level.

[00:20:40] They have not been able to come down to the assembly or the mid-market level, and it'll be a while before they come down to that level. That's fair. And Anurag, never apologize for bringing stuff up on this show. You are welcome to say anything you want because I bring you on for unfiltered opinions. So never apologize for you. I want to throw a question on from Steve, who's actually got a good one here.

[00:21:01] How, particularly to get your opinion on how can MSPs and IT service providers learn about performing AI assessments and workshop in order to create those projects? Like, do you have particular advice or insights on the way that process is being built? Oh, yes. So I think the importance. So it's a great question.

[00:21:21] The reason why I say that is a great question is because we did a study of, you know, over to 2000 partners where we looked at which what types of MSPs are able to scale their AI business and we call them the scaling partners versus those. Those MSPs were not able to scale their business. We call them the static partners, scale their business and AI specifically.

[00:21:45] And we found, you know, nine different areas where the scaling partners are doing a lot better than the static partners. And one of the areas was this building a readiness framework or the pathfinder plans. Right. And there are many MSPs who are really doing it well. And the way you have to kind of think about is, hey, you have to first invest in your own data management and data insights capabilities.

[00:22:12] Right. Because that's really part and parcel of that readiness assessment. Now, number two is you have to really invest in understanding what the customer's business is. What is the vertical? What is the customer business processes are, whether it's a retail or wholesale and things like that. Right. You have to have a complete deep understanding. Right. And then you need to have some kind of a services showcase.

[00:22:38] You know what that services showcase means that it goes beyond, you know, saying, hey, I can provide managed services, but it's about understanding as to what type of legacy systems they have got. Do I have the capabilities to really architect and deploy a solution before even I can manage that solution.

[00:22:59] If you have got those three capabilities, then you can sit down and either utilize your vendor partners framework tools or you can build your own. Right. There are many already in the market which have been branded by the MSPs. You know, whether you call that, you know, trade winds or tailwinds, I forget the exact names, but there are many such that. And by the way.

[00:23:26] Or for example, Lenovo calls it fast start or somebody else cost it, you know, quick start and things like that. You have to have creative. But the point here is that the SMB customer and the mid market customer are willing to pay for it. Agreed. And the other thing I'm finding is that data readiness is a good component of that. I had Colin Graves on the show, both on the live show and on an interview.

[00:23:52] And one of the things that he talked about the fact that they're doing is a lot of even before AI assessment, like a key component of that is data readiness, like doing data inventories, understanding access levels. So my advice that I've been giving to a lot is, is to go actually look at data as part of this to understand that because that is a key element of success with any AI project, particularly generative AI ones, where you have to know what your data sets are.

[00:24:19] By the way, for, for listeners, you're watching us do throw those questions in here. We got a little bit more time, but I would be remiss if I did not bring up the other area I wanted to talk about today, because you've written a lot about unified experience platforms as becoming a top priority for end customers. And before the show, we were even talking about how much that we think that there might be a real gap in the space of readiness. So talk to me first, like what are unified experience platforms and why are customers caring about?

[00:24:48] Right. So I think we should understand what the, what the background is, right? I mean, there is this, this traditional customer and employee experiences are often fragmented across disparate systems, whether it's CRM or ERP or marketing automation and HR and all that, right?

[00:25:06] A unified experience platform aims to break down the silos, providing a cohesive personalized and seamless journey across all touch points within, within and customer. Right. And what we are seeing is that I think the percentage which was there in that analysis was that 84% of the SMBs are not looking at just the customer experience platforms, or they're not looking at the employee experience platforms.

[00:25:35] Right. In fact, 84% of the SMBs are prioritizing both to streamline internal and external communications. And, you know, we are seeing that the convergence of contact center as a solution or as a service or unified communication, that service or unified communications, all those are actually interlinked. Right. And the reason why we call it the unified experience platforms is the fact that with this SMBs are not looking at the same time.

[00:26:05] We are creating some kind of blockchain. None of us can expect to streamline, I think I've said that is streamline internal and external communication, fostering a more connected workforce through improved team collaboration increase productivity. At the same time, enhancing the customer interactions, delivering really superior customer service, which drives customer satisfaction and brings stronger relationships. It's not either or situation.

[00:26:34] If you have to drive a competitive edge, simplify operations, increase efficiency, and thrive in a dynamic market, you need to have both. And both of these need to be really integrated. And the reason why, because UXP in our thinking turns the fragmented journeys into fluid experiences, right?

[00:27:00] And from data chaos to experience orchestration is the key aspect of it, right? It's the UXP platform. The way I see it is the conductor of that customer and employee delight. Now, I want to make sure that I'm understanding this right, because there's a lot of concepts you're trying to pull together. It sounds like the idea is, traditionally, we've thought about like a customer relationship management. So I manage my customers here, where I manage my employees

[00:27:28] in like an HR management system. And then I manage my online experience with like my website and my e-commerce platform. And it's the idea behind a unified experience platform that what we want to do is we actually want to pull these systems together so that we're managing all of our interactions with all of our stakeholders. I'll use this as a broad, like customers, partners, vendors, employees. Is that the idea that you're trying to bring it together into a single vision? A hundred percent. Okay. Absolutely right.

[00:27:57] Because we are in the experience economy, right? And in today's market, experience is the key differentiator, right? And UXPs would be the technological backbone for delivering this. Now, I want to make sure that I like distinguish this because this runs the risk of being really good marketing, right? Is there a set of sort of key features that when you think about

[00:28:26] a unified experience platform that would set it apart from traditional CRM tools or collaboration tools? I think it really starts with the communication vendors. Okay. Right? Whether you look at a NICE or a 5.9, you know, VooSever or TalkDesk and others, right? Or Genesis, RingCenter and all. They actually tend to focus on either or situation and then they come out

[00:28:55] and tout the fact that they have got a hundred different integrations or 80 different integrations or 10 different integrations, right? Integration brings in complexity, right? But when you really have to deliver a truly unified and personalized experience, you have to have a significant amount of customization. Partners can build that, right?

[00:29:24] But then if there's a communication vendor comes out with, on a unified data platform, both employee experience and customer experience, that makes it easy, right? Because then it drives operational efficiency and data harmony. Because by consolidating the data and processes on that unified data platform, the UXPs reduce

[00:29:52] the redundant efforts, improve data quality, provide better insights, leading to increased operational efficiency efficiency and reduced costs. Now, how much of this is theory and how much of this is stuff that you're actually seeing in the market right now? Because I like directionally where this is going, but I also feel like there's a lot of thinking of the direction versus what's in the market. Are there particular solutions that are really embracing this now? It is happening

[00:30:22] in the market, right? I mean, Zoom is approaching this. They don't call it the UXP. They don't call it a unified experience platform. Everybody has come out with their own nomenclature. Then, you know, you call it about Cisco and before the call, we were discussing about Cisco as well. So Cisco, the whole WebEx suite is capable of doing it. It is just that nobody has kind of put it in this framework that, you know what?

[00:30:51] The CX and the EX platforms are one and the same thing. Look at you bringing up WebEx. I appreciate that. I appreciate that. Now, I also have to sort of say, like, it feels like there's a real gap in readiness between what a lot of the, a lot of this discussion, if we think about, like, you know, Zoom or WebEx or like where they're investing, it feels like there's a gap in readiness on the partner channel side to be ready to actually put this into place

[00:31:20] for most customers. Am I right to think of it that way? And like, how large is that gap? Absolutely right. Because see, what the MSP channel, the channel does is that they tend to gravitate towards what the vendor partners, marketing, and selling points are. Right? And this is one of the things that we are advising the MSPs is that you have to move away from, from vendor push

[00:31:51] to customer pull. Right. See what the customer is asking for because your loyalty relies with the customer, not with the vendor. And then look at what the customer is asking for and create your own capabilities, right? And in, in some cases, it's going to be a low-hanging fruit for the MSPs, right? Because this is a multi-vendor integration play, right? Can you just combine capabilities of different vendors, right?

[00:32:21] You know, you can be that essential component in connecting legacy systems or existing systems with new UXP components, managing their data migration, ensuring data integrity across the unified platform. I think there is tremendous amount of capability there. You know, and you've said it in a nicer way, I'll say it a little bit stronger. Like my biggest frustration of most channel discussions is it seems to leave the customer out of it entirely. It talks so much

[00:32:51] about the vendor to the channel, but it forgets that the customer has to actually be interested in that stuff. But I've got another question from Steve, and I want to make sure that we bring it, particularly because you've got, potentially have some insight into this. Do you have any sense of the market, like percentages or market adoption around self-service applications? Do you have any kind of sense of the way that those, you know, how much self-service being delivered for clients? You know, that question came up recently

[00:33:20] from one of our clients. I do not know what we responded to that client. We don't have the data per se, but this is where a lot of the vendors and the channel partners are also moving towards, you know, as to how they can enable that self-service capabilities. We are seeing a lot more within the SMB segment in the category of, say, 100 employee size

[00:33:48] to about 1,000 employee size, what we call as the core mid-market. But that is also where the new paradigm of customer success is actually moving towards. And there are many different vendors. So the short answer to the question is, I don't think we have the data. If we have the data, I don't remember the data at the top of my head. But this is where a lot of different vendors

[00:34:17] are investing in driving their customer success solution with the self-serve capabilities to drive down the number of tech support calls that they get. and they are using agentic AI capabilities. I have been part of many different briefings over the last three months or so that this is where the investment is actually flowing in.

[00:34:47] I think that's going to be, and I think, Steve, to a certain degree, we may not get a lot of insight into it because in some level, you know, the MSPs that are delivering this kind of solution are not necessarily going to report that in an obvious way that we as outsiders are able to see on their organization. I don't know, I want to be, go ahead. So there was a point I really want to make because you bring up an interesting point because one of the things that we found when we did the study on the scaling partners with the static partners, what we found

[00:35:16] that the scaling MSPs are really investing a lot in their own AI capabilities. And I think there was a point at the step number seven or step number eight, it was where they're saying, you know what, how do I improve my own service and delivery model? And many of these progressive scaling MSPs are testing their own champagne. That means they're deploying AI to really understand what their travel tickets have been over the last 10 years or 12 years or 15 years

[00:35:46] and harnessing that information and then using AI capabilities and building, you know, self-serve capabilities for their end customers. I think you're 100% right and in fact, some of the conversations that I've had with a few MSPs exactly do into that. For listeners watching to this, look for the interview with the team from NetSoret who was on the show and talked a little bit about the way that they're building chatbots for their own organization and I've been tracking several conversations exactly to what you just brought up where they're doing

[00:36:15] ticket processing, not necessarily resolution of ticket but making sure that the information coming in from the customer is organized in a standard way, that they're pre-populating potential solutions to issues, you know, as they come in. I think that those are some of the things that we're seeing out there. But I want to be respectful of your time as we wrap up our conversation here and throw any last questions in to wrap up our time. But my big one to you is like as you sort of are looking to, you know,

[00:36:44] we're now in the second half, back at half of the year. What are you looking for as kind of those big trends or something you're excited to see potentially happen before we get to the holidays? Before we get to the holidays. Right. You know, I think the three topics which you brought up, right? Let's talk about AI operating system. I think one of the trends that I'm really trying to understand is,

[00:37:15] you know, can there be a time where the OS does not just run apps? Can it think for them? Right. Can it think for those apps? Right. And that requires a contextual orchestration layer, right, for the evolving interface between AI and OS. How soon will that happen? Right. That's one. Number two is what is going to happen

[00:37:44] in terms of, you talked about AI security. What is going to happen in terms of cognitive cyber defense? How can AI be leveraged for smarter security? I think that is going to be important, right? Everybody can talk about platform or point solutions and things like that, right? And then the third area you talked about was the UXP, the unified experience platform, right? Can UXP's role

[00:38:14] be akin to a journey harmonizer, right? I am on a journey, right? Can it create the seamless interactions between the employees within the employees, then employees and the customers? I think those are the three areas that I'm really looking forward to. You always bring such interesting ideas. I'm going to offer a premise a little bit on the first one. I think the first one's further out than we both would want it to be.

[00:38:44] I would sort of say, if I look at what's happening right now with Apple, if I'm looking at operating system vendors, I essentially have three to look at. Apple, clearly distinct problems in their AI rollouts. They're having all kinds of personnel churn, Meta's diving in there right now and trying to steal some of their people. I think that they are a significant distance away from having something useful. Google is leaning much into the communication and processing side of it, but I think, if we think about them from an Android

[00:39:13] or a Chrome perspective, there's pieces of it, but I definitely wouldn't see it this year. Microsoft is interesting in that they have a lot of the pieces and particularly their NL web components and their rollout of MCP had some really interesting components to it, but at the same time are heavily reliant on partners, i.e. OpenAI, who continues to set money on fire without actually being profitable. And so there's enough of the components

[00:39:42] that I see your vision. I'm not sure it would happen in the next six months, but I'm looking forward to having you back to have the conversations. Do it. I don't know. A lot of fun with you. Always a lot of fun. If people are interested in reaching out and getting in touch, what's the best way for them to do that? Yeah. Reach out directly at my email address, anurag.techisle.com. I respond to each email. And listeners,

[00:40:12] he's really great about getting back to that. We are always having fun conversations. Anurag, thanks so much for joining me today. Thank you for having me. It was such a blast. Now, I want to tease an interview I'm releasing this weekend. Bob Burke joined me. He's the chief information security officer at Beyond Identity. And he challenged the effectiveness of traditional multi-factor authentication in the conversation. And he talked a lot about the need for phishing resistant solutions. We had a conversation also about his FedRAMP journey and what they could learn.

[00:40:42] And I want to give you a preview of the conversation we had and give you some of his thoughts on the SSO tax and what's going on in the market. Here's a preview of that interview. There's a lot of conversation around the SSO tax, right? That many companies, particularly software as a solution, liver solutions, put proper integration of single sign-on behind a much higher tier and they make customers pay tier to have that. So, for example,

[00:41:12] and it's significantly more expensive than most products. You know, what's your position on the way SaaS should be approaching this? Because right now they're actively discouraging customers from being more secure. Oh, that's exactly right. No, that's exactly right. So, if you look at some of the legacy solutions and like I said, they're all great companies with great products and great teams, but they're struggling with this Q sprawl based on a legacy solution. And so, when you start looking at it, and this goes back

[00:41:42] to the core proposition of any solution that someone needs to buy is it has to be secure by design from the ground up. You can't actually look at security as something you buy post the SSO and that also applies to compliance. It really has to come and it has to be part of the original solution and it needs to be cost affordable. You can't expect someone to buy an SSO and then through 15 different SKUs add in the appropriate level of security that you need.

[00:42:11] Certainly when it comes to, maybe you can argue there may be a really sophisticated ITDR solution, IPSM or something, but when it comes to fishing resistant authentication, that is your authentication layers go to the business and you should expect that from the, whatever the SKU is, that should just come right out of the box. Bob, I had some really great positions that I think MSPs are going to be interested to hear and my Patreons already have this interview now if you want to get access early.

[00:42:40] It's a benefit of supporting the show. Visit patreon.com slash MSB radio to sign up and get access right now. The interview will drop on the YouTube channel and podcast feed this weekend. Now I want to thank Sales Builder, our Patreon sponsor whose support makes this show possible. Focus on your IT sales workflow with the power of automation and visit them at salesbuilder.com that's B-U-I-L-D-R dot com. And vendors, you too can get your name mentioned on the live show. It's a simple monthly subscription.

[00:43:10] Visit patreon.com slash MSB radio to learn more. And listeners, here's what you can do to help me out. You can support the show directly. Make sure to like, share, and follow on your favorite platforms. Hit that like button. It makes a big difference. If you want to support directly on Patreon and get content early, you set what you think the content is worth with our give what you want model. Sign up at patreon.com slash MSB radio. If you have a question

[00:43:39] and are listening to the recording, either put it in the comments below or send it in at question at MSB radio.com and answer it next week on the show. Thanks for joining me for the Business of Tech Lounge and I will see you next time. I'll see you next time.