Agentic AI and Automation: Transforming SMB Operations with Anurag Agrawal of TechAisle

[00:00:12] What trends must SMBs and mid-market companies watch for in 2025 and how can partners capitalize on them? Anurag Agrawal of TechAisle joins me to explore the rise of agentic AI, the impact of AI powered PCs, and how SMBs can tackle top business goals, IT challenges, and their technology priorities. Welcome to the Business of Tech Lounge, the live version of the Business of Tech podcast. It's Wednesday, January 22nd, 2025.

[00:00:42] And I'm Dave Solberg. I'll be taking questions and comments throughout the show, so make sure to put them in chat. If you have a question, we'll happily respond to them. 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.

[00:01:07] Now, reminder, I am keeping an eye on that chat, so we'll be watching for your questions and your comments throughout the show. Anurag Agrawal is the Chief Global Analyst and CEO of TechAisle, a global SMB and mid-market IT market research and industry analyst organization. He's got nearly 40 years of experience in the industry, and he's recognized for his expertise in the SMB market and channel segments. And on top of that, he's a good friend to the show. Anurag Agrawal is the Chief Global Analyst, and he's a good friend of mine, and he's a good friend to the show.

[00:01:37] And I'm glad to be here. So, Greg, welcome back. Happy to be here, Dave. Thank you for the opportunity. You know, I always love having you on, and I think we're going to make this a bit of a quarterly thing because you always come ready to play with some of the most interesting projection research. And I want to talk a little bit about your 2025 SMB and mid-market predictions, because you've made a couple of bold ones for this year. Give me a sense of kind of like the most one or two significant trends that you're keeping an eye on for this year.

[00:02:08] So it's about AI and AI and AI, right? It's about, you know, last year was about experimentation. This year, it's about ROI, right? And the SMBs are saying, you know what, we have done a whole bunch of experimentation.

[00:02:22] We have deployed many, you know, POCs, but we really want to see how this AI can really drive business success because of the fact that many of these SMBs are also clamoring or really fighting for that talent which is out there in the marketplace.

[00:02:43] So they are turning to technology like AI to say, you know what, how they can automate these processes. So there are three things we are looking at. One is the rise of this vertical AI, such as vertical solutions for whether it's retail or healthcare or financial services or even professional services organization. How can some vendor or partner can come up with and deploy vertical AI solution?

[00:03:09] The second is about automation and automation in a box. What that means is, you know what, there are business processes that SMBs have. They may not be as complex as an enterprise customer, but they still have complex business processes. How they can automate that. How they can automate that using agentic AI because that sounds like a very promising technology because SMBs adopted RPA. Now they're saying, you know, how they can take it to the next level.

[00:03:39] The third one is about AI PCs keeping with the same theme of AI. Yes, the adoption level is low. You know, only about 19% of the PCs likely will be purchased. By SMBs will likely be AI PCs, but the year has just started. So we are looking at those three things which actually percolate to the top in terms of the trends.

[00:04:07] Now I'm going to come back to AI PCs because that's what I'm interested in, but I'm going to dive in a little bit here on agentic AI, because obviously this has been one of those themes that gets a lot of buzz, right? If I'm one of those media consumers that I'm looking at the news all the time and there's constant discussion about agentic AI. But at some level, this is still a little bit of promise more than it is actual in the field technology.

[00:04:30] Particularly when I'm looking at some of the big providers, if I'm picking on open AI or anthropic and the way they're approaching it, they're still kind of coming up with the way this is going to work. How much of this are you looking and saying is this projection of what's going to happen and how much is actually happening in the field right now?

[00:04:46] So, honestly, what we are trying to say to the SMB AI wave, right, even if it's considered to be high by many, right, don't lose out on this wave.

[00:05:02] They were late to adopt RPS solutions. But when the SMBs for the last five years have been looking at this business process automation and how they can turn to ERP solutions or CRM solutions or, you know, invoicing solutions at the low end, or whatever types of SaaS based applications. Hey, how do I employ an IT consultant or a developer in house to really write that integration code using, you know, a bunch of APIs?

[00:05:31] Or I should look at agentic AI. I think that's what we are trying to tell the SMBs in kind of building this out. So, in a way, yes, it may be hype, but it is going to be very real. If you talk to, you know, Anthropic or you talk to OpenAI or talk to Microsoft or, you know, you have Salesforce, which has come out very strong right out of the gate, they're all talking about this agentic AI.

[00:05:57] So, this is the year of the AI. So, in a way, what we are trying to say, and I'm repeating myself because it's a point that I really want to make, is that this is a capability that they should not really ignore, right?

[00:06:13] Because this is the capacity to automate the workflows. But it's not going to be easy, right? Because they have to hire either a consultant or they have to have an internal champion who must invest in process analysis, as well as employee upskilling.

[00:06:31] And what we are going to see is when we talk to the SMB customers, many of these progressive SMBs are really becoming the early adopters because they are focusing on augmenting their existing workflows, the workflows where automation is currently failing for them.

[00:06:49] So, how much of this is early and what they can work on? Because, I mean, I get it. Like, I want customers to be thinking about that. And ideally, I want partners to be spending time investing there to get their customers ready. But we also, at the same time, want to sell things that are in market, right? That we want to actually take solutions to customers that they can do something with.

[00:07:09] We want to get the timing right on that. Are there particular sets of guidance that you can give to partners to say, these are the things I want you to be thinking about and working on now, while we're as a preparation for when those products come into field? What a great question. I was not prepared for that question. But let me answer that.

[00:07:32] So, in a study that we did, 65% of the SMBs are aware of our agent TKI today, which is very important to understand that. Right? I mean, the SMBs range from, you know, less than 10 employee-sized firms to about 1,000 employee-sized firms, right?

[00:07:54] There is 42% of them are interested in this agent TKI, which is a pretty high number. Think about that for a minute, right? But on the flip side, only 42% of partners are really offering AI solutions to their end customer. They have not even reached at the agent TKI level. Right? So, there's a huge gap between what the SMBs want and what the partners are offering.

[00:08:22] Now, out of those 42% of the partners that are offering, you know, AI solutions, only 11% are able to kind of scale, right? In offering AI solutions to their end customers, right? So, the answer to your question is what they need to do. First of all, the most important thing they need to do is to look at their own offering from a customer-in perspective rather than their product-out perspective.

[00:08:51] They have a tendency to sell their products, their capabilities that they have got in terms of SaaS solutions, in terms of other capabilities, but they have to really invest in building a consulting practice, which means that they need to invest in having a business process person within their organization. Now, this is like really the meat of the conversation, right? So, I'm really curious.

[00:09:19] Do you have any particular insight on why that gap is forming? Because if I think about it, like it feels like the customer is further ahead than many of their solution providers. That's not often the case, particularly if I think about this, like if I think about cybersecurity, right? Because if I think about cybersecurity, we're often pushing the customer to do something and they're resisting. Here seems to be a case where the customer is much further ahead. Why?

[00:09:46] So, if you go back more than a decade, right? Maybe a 2008, 2009 timeframe, right? Cloud was like that because cloud was as an agility platform. It was touted as adding to the top line growth of an SMB customer. And the SMB says, you know, I want to adopt cloud. But the partners came along. They were slow to really offer cloud to their end customers.

[00:10:15] And the biggest excuse was our customers are not ready for it. Because it's not about being ready. It's not. It was. The problem was they were trying to lead with the price point. They were not showing the value capability of what cloud could do for their SMB customer, right? So, the entire history is actually repeating itself here.

[00:10:39] Now, what is also happening is that partners themselves are fighting for that AI talent. And how many AI-scale people really want to work for the partners? I mean, that's the truth. That is the reality, right? And then these partners are also being bombarded, if I can use the word, by all their vendor partners to deploy AI-infused applications.

[00:11:09] But the customers are saying, you know what? I know I can get, you know, SaaS applications, which has got AI capabilities. But if you understand my business process, how do I do that? So, in a way, the partners are not investing in their consulting practice. They are not investing in their skill sets. Skill sets means, hey, you know what? I need to have data management capabilities. I need to have data science capabilities, right?

[00:11:35] I need to be able to have data streaming capabilities, right? So, they're not investing in those areas. They are actually gravitating towards just deploying, say, co-pilot, which is okay, which is not bad. But then AI is much broader than deploying co-pilot. So, that may be some of the key insight there is too many partners are looking at it in terms of deploying a product or some within the SaaS and not enough on the actual solutions that can be built around AI.

[00:12:04] Now, I'm curious, this traditionally has not been a space where the partner is doing a lot of development. Now, I want to acknowledge that in an AI-powered world, not all of this is writing code, but some of it is at least thematically very close to product development and custom code versus their normal practice of bringing in applications and technology solutions and just kind of assembling them.

[00:12:32] Is that a key component of the skills gap that they're having to deal with? And if so, like how big a factor is that in all of this? It is a pretty huge factor because, you know, the partner should actually think back and say, you know what, if I am able to put in my resources in building my competency and specialization instead of chasing certification from the vendors,

[00:13:01] I think that will be a very key important first step. What that specialization means, hey, you know what, I need to have some specialization. Maybe it is just a simple thing as deploying voice assistance, right? Let's start with that area, right? So, understand as to what your capabilities are. Maybe work within one SMB marquee customer, even an enterprise customer. I think that launches your capabilities and your skill set.

[00:13:30] But more importantly, we study 2,400 channel partners, you know, saying, okay, you know what? In this new world, what is separating the partners that are investing or they are scaling in deploying AI solutions for their end customer, regardless of whether they are SMBs or enterprise customers, you know, one important thing is the partners themselves have started to use AI internally. Interesting. So, that's another bit.

[00:13:59] Now, I want to talk a little bit about the AI-powered PCs component because it's interesting to think about AI-powered PCs in the year 2025 because we're additionally going to be in a Windows 11 upgrade cycle, Pierce. And Microsoft is clearly heavily invested in making sure that that upgrade cycle goes very well. As I've just reported on the show, not only are they pushing the end of life for Windows 10,

[00:14:25] they've now coupled Office and Microsoft 365 as also being end of life on Windows 10 at that exact time. So, they're clearly invested in making this happen. How much of this upgrade cycle are you looking at as being powered by their drive on the Windows upgrade side? And how much of this do you think is going to be related to the capabilities of AI-powered PCs? That's an interesting question, too.

[00:14:51] So, we did a research and we found that, you know, that 77% of the businesses are actually aware of implications of the Windows 10 end of life. So, we've also shown that a lot of these people are not able to be able to get the same support very clearly, right? Out of which, if you look at just from an SMB perspective, 21% are not swayed by the fact that, you know, Windows end of support will come.

[00:15:15] They say, you know, their channel partner will continue to support or some of them are even willing to pay, you know, $61, you know, for an ESU, right? to say, you know what, let us extend it till we are ready. But there is an interest. So now, very clearly, what is happening is that these end customers,

[00:15:41] specifically SMBs, are actually separating into two and saying, okay, you know what, should I buy a co-pilot plus enabled PC, which may not be a typical AI PC by the definition of it, in terms of the number of tops and all that sort of thing, or should I go in for an AI PC? So that's where it starts to become very interesting. Because yes, you know, they want to buy, they want to refresh new PCs.

[00:16:11] Yes, absolutely. Right? Those who are saying, hey, end of support is coming. They're also thinking that 29% of the PCs that they may purchase may be enough by just buying a PC which is co-pilot enabled, which means that they've got a co-pilot plus key, and then they can pay Microsoft separately for using co-pilot, which in that case, they do not require an AI PC.

[00:16:39] Now about 14% of them, or 14% of the PCs likely may be AI PCs. But that's where it gets very interesting. Because when you start to think about their entire AI stack, right, which may think, hey, let's start from the infrastructure level going up to the software and services level, 50% or actually 47% of the SMBs kind of think that AI PC could be relevant as the core AI foundation.

[00:17:06] But at the same time, they don't know as yet as to how AI PC will create that differentiation within their entire organization. So there is a lot of education that needs to be done, both from the channel partner side and from the AI PC vendor side to really outline the advantages of using AI PCs as these SMB customers really accelerate their AI adoption.

[00:17:35] Now, is that what you've targeted on the education portion? And by the way, I will generally observe that most channel problems can be solved by good education. And what I want to get a sense though from you as the analyst looking at this, is your sense that it is purely that they've not done a great job on the education or is the business value still TBD? Because there's a good portion of this where it feels like there's a lot

[00:18:00] of AI hype without necessarily a lot of measurable ROI on those outcomes. We've seen a lot of pilot projects in the past 12 to 18 months, but not as many of those emerging from pilot into full production. And so I'm trying to get a sense of like, do we have a core problem here or is it purely just education? Well, most analysts will talk about the fact that yes, there is a lack of use cases. It is true, right? There's a lack of use cases.

[00:18:28] I mean, yes, I mean, I have got like three or four different AI PCs at home. Do I use them for AI capabilities? Yet, not yet, right? But the point here is that if you look at the top three PC OEMs, they're coming out with some good offerings, right? For example, you have Dell coming out with this AI, what they call as the Pro AI Studio, you know, Dell AI Studio, which is kind of, hey, it is fine-tuned for different models

[00:18:55] that an SMB customer, for that matter, any customer can use locally on their PC. Then you have, you know, Lenovo. They have something called the AI Now, right? Which is, you know, it's curated, validated models. And it does, you know, day-to-day tasks for an employee or a user on a PC itself. Then you have, you know, HP. HP, I think they call it the HP AI companion, right?

[00:19:24] I mean, they came out of the gate first, maybe last year. They kind of announced it. So they're all trying to work, come up with some AI use cases, make that AI PC a lot more useful. But then if you start to think about it and say, you know what? There are IT departments. There are creators. There are sales marketing folks and all that. They're all experimenting with it and say, you know what? If I am doing some kind of a, you know, on-site customer support, for example,

[00:19:53] or, you know, customer service or doing some sales calls, can I use an AI PC instead of connecting to the cloud to kind of drive certain features and functionalities? For example, an architecture company, which I was speaking to last week, they said, you know what? With the AI PC, they have got this whole MetaLlama model, which is running locally on their PCs.

[00:20:23] And they are able to kind of develop that architecture, the remodeling of the house of an end customer locally without connecting to the cloud. And we are able to kind of do that very, very quickly. So there are some use cases which are actually coming up, but it's opening up tremendous opportunities for the partners themselves. Now, I mean, again, this is why I think this is interesting, but I want to make sure that I'm looking at this from the perspective of the partners that can excel. One of the things that's interesting to me is I think that this is going to be one of those spaces

[00:20:53] where we separate haves from have-nots a lot. I think that what we're seeing is we're starting to see a generation of providers that are pulling ahead, that are actually being good at this, and widening the gap on their effectiveness, and thus also their profitability and their engagement. Are there particular traits that you're looking for in partners that can identify those higher-performing partners, particularly as it relates to AI?

[00:21:23] Oh, there are, I mean, that can be a separate... But, you know, we are looking at, we have identified nine different steps based on the research that we have done, right, which really highlights as to what partners can actually do to really be successful in deploying AI solutions for their end customers. I'll give you a flavor, right?

[00:21:51] For example, when you start to look at areas around, say, you know, what are the foundational capabilities and knowledge, right? First of all, they need to build the AI data foundations. I talked about data management, data science, data strategy capabilities. Then they need to create a services showcase. What that services showcase is, you know, that they are actually leading from the services side rather than from the product side.

[00:22:21] Then you talk about the customer engagement and understanding area. It is very, very important. Prioritize deep customer understanding from the start. That means even before you contact the end customer. That means you need to use AI capabilities to understand what your customer has got, right? Then you talk about the focus on value-driven customer relationships. And this is where many of the partners are actually changing how they approach, right?

[00:22:50] For the longest period of time, the partner started using the sales approach, unknowingly called the band approach, which is, you know, let us identify what the budget the customer has got, what the authority is, and so on and so forth, right? But now the partners, 41% of the partners are kind of moving towards the champ approach. We're saying, you know what? Let us first spend 39% of our time in identifying what the challenges are of the end customer.

[00:23:17] Then let us identify who is the authority, which is the A part of it. Authority means the decision makers. And we are finding that typically in the AI, there are 5.6 individual decision makers in an end customer. So if you are talking to IT only, you are on the wrong side of AI, right? You need to identify 5.6 decision makers. And then we can kind of go on and go on.

[00:23:43] But the more important thing is that most of the partners have shifted to account-based marketing and high-touch support. Interesting. I want to know how to find 0.6 of a person, though, because I think that'll make the difference. But I want to pivot a little bit because I don't want to spend the entire time on AI, because one of the other areas that you have talked about is cybersecurity, right? And what I want to get a little bit of a sense of is while everyone talks about cybersecurity being hot, I'm starting to feel like this is a little bit more of the same.

[00:24:12] Is there something different in 2025 that we should be thinking about from a cybersecurity perspective, or is it more just we have to keep fighting the cybersecurity battle? Give me a little bit of your thinking on cyber. No, that's another great question, Dave. You know, that's why I love coming onto your podcast and all. See, the interesting point here is that cybersecurity has the top-of-the-mind IT challenge for the longest period of time, right?

[00:24:40] But has it, are the people really spending, are the organizations really spending? Is the spending on cybersecurity going beyond 10% year on year? Right. Right. And the question is no, right? It is still stuck, right? Because of the fact that, hey, you know what? Where should I start? What do I do? What is the safe enough? What is the good enough capabilities, right?

[00:25:10] And now you have this AI coming into the picture. Can I use cybersecurity AI capabilities to fight AI-infused, you know, cyber attacks? Those are the kind of questions which are coming up. But what we are seeing is, again, within the SMB context, you start to see, hey, they want integrated solutions. They are actually tired of these point solutions, right? They say, hey, you know what?

[00:25:38] And Palo Alto Networks has famously talked about this fact, you know, Nikesh Arora about platformization. We have been seeing that within the SMB customer for the last two years, right? But then no single vendor is really focused on the SMB customer. They say, hey, you know what? Yes, there are vendors which are focused on it, but they, you know what, outside of selling point solutions, how can they really build some kind of a cyber strategy for an SMB customer?

[00:26:08] Give them the good, better, best option. And the channel partners themselves are limited by their own capabilities as to how many vendor partnerships they have got, right? They can only, you know, they have the hammer and nail problem.

[00:26:20] So, you know, I'm not saying much, I'm saying a lot without saying too much, but the point here is the fact that, hey, things like identity access management, SIM and SOAR, you know, these are also important for the SMB customer. Who is actually addressing the needs of the SMB customer?

[00:26:43] If you think about the data sets of the last two years, they have always been on top of the list of that SMB customer that we want to adopt it. But do you have the capabilities, the products, the solutions to really deploy that? Because for the last two years you've spent, the vendors and the partners have spent in educating the SMBs about EDR and MDR and XDR. Well, the adoption is still not 100%, right? Right.

[00:27:13] I want to get your thought on a little bit of something that I've been giving a lot of thought to recently. And you're somebody who thinks a lot more about the vendors' engagement here than I do. So I really appreciate you. And by the way, I may be completely off base here. That's kind of the fun of it. Is I think most SMB customers generally want cybersecurity problem just solved for them, right? They want a partner that come in back. And I think actually one of the underutilized areas that this could be done is it's actually fully in the network layer.

[00:27:41] I think that many small businesses would be happy to pay their connectivity provider to give them a safer version of the internet. For example, I always laugh and go, we roll up trucks with this internet line that is just gross. It's not clean. It's just all of the grossness of the internet. You get the whole thing. And I think there's a lot of businesses that would just say, I want one that just doesn't have all of that stuff. I want it to be a little cleaner.

[00:28:09] There are probably countries I don't need to talk to. There are portions of the internet I don't need to talk to. If those people want to do it on their own time, they can go home and do it there. I think they should get a safer version of the internet. How much are vendors thinking about just fully integrating that into the network? Or am I just totally off base on my thinking here? I think you are on the right track.

[00:28:31] But it's an industry's endemic problem where you have 4,500 security vendors, right? Who can rise to the top where nobody has more than 5% of share or 10% of share of the security market, right? How do you actually rise to the top, right? And I think the way the industry is moving is, yes, the SMB wants that, right?

[00:28:57] How do I make it safe for my data, for my users, and for my devices, right? How can I use the principles of zero trust within my organization, right? So organizations like SentinelOne have started to think about it. Organizations like Lenovo have started to think about it, right?

[00:29:27] Hey, can we start to build that in right from the device to the network to the cloud, right? Can I build that? And then can I have cybersecurity embedded in the entire AI supply chain, AI solution supply chain? So their vendors have started to kind of think about it. But the reality is that most of these vendors focus on the enterprise first, then they come to the SMBs later. That's so true. We see a lot of that.

[00:29:56] So a last sort of thought here is give a little bit – we both are – we love the SMB in the mid-market space. Yeah. Let's do something a little bit basic to help partners think about it. When you look at these vendors and you look at the various solutions, what's kind of the key indicator for you that a vendor's strategy understands the SMB in mid-market space? Yeah.

[00:30:20] So when they lead with not a price point because they say, you know what, vendors, we understand. So there are two things there. One is they don't want to lead by telling the SMBO, we know that price is important for you. Okay. Or they don't want to lead with the fact that, you know what, IT staffing, we know you have got less IT staffing.

[00:30:46] Well, you know, these two statements have been bandied about for the last 30 years. Right? SMBs get it. The vendors get it. Right? So when a vendor says, you know what, we understand that these are your three business challenges. These are your three technology challenges.

[00:31:09] And how I am going to address these with the solution that I'm offering you is going to be very key. So that's my litmus test. That means a vendor is able to connect the business and IT challenges to the value map of the product solutions or the offerings that they are offering to the SMB customer. So I think that is music to my ears. Well, I like a little bit of music. And so let's end it right there.

[00:31:38] Anoreg, I love having you on the show. We always have such great, fun conversations. And I cite your research all the time on The Daily Pod. But if people are interested in getting your research directly, how can they do that? They can go to techis.com or send us an email, inquiry at techis.com. Well, he makes it incredibly easy, listeners. And he and I will be having more of these conversations. Anoreg, thanks so much for joining me today. Thank you for having me.

[00:32:07] It was a lot of fun. Happy to do it again. We will do it again soon. Well, I have another highlighted conversation to talk about because coming up this weekend, I talked with Anil Gupta of YBot to discuss the transformative shift from mobile device management to declarative device management and what its implications are for enterprise device management. Now, he emphasized how this evolution enhances scalability, security, and efficiency, particularly around Apple devices within organizations.

[00:32:37] Here's a preview of that conversation. Yeah. So this all comes down to actually scale. You know, so as you see, you know, of course, everybody's using lots of Apple devices throughout, you know, in their organizations. And as the number of devices grow, the dependency on a single server to be able to manage all of that just kind of grows out of bounds, you know.

[00:33:04] And this paradigm actually applies to almost any IoT space, if you will, wherein if you look at the way kind of industry moved, you know, way back, there used to be, you know, the central kind of servers and dump terminals. If you go way back, you know, 20, 30, 40 years ago, you know, and then the paradigm of individual compute devices, laptops, MacBooks, et cetera, emerged, right?

[00:33:32] So people figured out that centralized servers eventually can become a bottleneck, and then it should transfer the decision-making knowledge, you know, all of that to an individual device. Now, my Patreon supporters already have this if you want to listen now. It'll drop on the weekend on the YouTube and the podcast feed. If you're interested, I really encourage you to join in and listen. Visit it at patreon.com slash MSB radio.

[00:34:00] Now, I want to thank Sales Builder, who's 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. Visit patreon.com slash MSP radio for more. And listeners, here's how you can support the show. Like, share, and follow on your favorite platforms.

[00:34:31] It really makes a big difference, particularly if you're following or subscribing. And if you want to support directly on Patreon, you can use our Give What You Want model. You set what you think the content is worth, and you get access to my video interviews early. If you have a question and are listening to the recording, send it in at question at MSP radio dot com. Thanks for joining me for the Business Tech Lounge, and I will see you next time.