Navigating AI Adoption and Governance in Technology Services with Seth Robinson

[00:00:00] Artificial intelligence, it's all anyone can talk about. But where is the value for providers? Where do you focus to create value for customers? Seth Robinson, Vice President, Industry Research at CompTIA joins me today. Welcome to the Business of Tech Lounge,

[00:00:31] the live version of the Business of Tech podcast. It's Wednesday, June 12th, 2024. And I'm Dave Sobel. Now we'll be taking questions and comments throughout the show. So make sure to put them in chat. There is a dedicated question section at the end, but today, particularly with the format,

[00:00:47] we're going to take questions and comments anytime. So make sure to just throw them in the chat. We'll throw them on screen. 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

[00:01:02] and visit them at salesbuilder.com. That's B-U-I-L-D-R dot com. And a reminder, I am watching that chat even when I flub up. Now our top story of the day. Let's get some context around this AI conversation. Last week, I reported some data on Barracuda

[00:01:24] about positioning and MSP thinking on AI. Let's dig into that data with this clip. According to Barracuda's global MSP report, 87% of MSPs feel the need to improve their knowledge and application of AI products and services to meet customer needs. IT service providers anticipate that 44% of their revenue

[00:01:45] will come from recurring managed services in 2024, up from 34% in 2023. They plan to introduce an average of six new services, primarily security related. The report also revealed that 77% of managed services providers feel under pressure to provide AI insights and tools to their customer. So they're feeling pressure.

[00:02:08] There's pressure to provide insights and tools to customers. So what does that mean? Joining me today to discuss is Seth Robinson. He's the Vice President, Industry Research at CompTIA. Seth, thanks for joining me today. Hi, Dave. It's a pleasure to be here.

[00:02:24] So good to see you again. Thanks for inviting me. Oh, it's great fun. You've been a frequent flyer on the Business Detect show, and I've been on your podcast a couple of times, Volley, which you can catch on all the podcatchers of your choice.

[00:02:35] And you guys are on YouTube now too, right? Yep. We're on CompTIA's main YouTube channel. So if you want to see us doing our thing, you can check us out there. Awesome. So you were the guy I wanted to talk to on this,

[00:02:46] particularly because I knew that we could have an interesting conversation, but the timing is great because you've actually been working on a white paper and research around how providers can think about this. Lay out your thinking right now. I've spouted mine a bunch of times on the show

[00:03:02] in the way that I think about this, but give me your thinking on the way solution providers right now can get some value out of this AI conversation. Yeah. I mean, I think that it starts, Dave, with a conversation that you and I have been having

[00:03:16] for several years now around emerging technology, both within some of the work at CompTIA, some other things that you and I have talked about. We've been having this discussion around emerging technology, and I think one of the main themes of that discussion

[00:03:30] is that most of the emerging technology that we're talking about is not a product. It doesn't have a SKU. It's not something that comes in a box. We've talked a lot at CompTIA about strategic IT and the shift in energy that's been happening

[00:03:45] from putting a lot of work and investment into the computing platform and starting to put that work and investment into applications. And so applications are getting more and more complex, and you're getting more layers in that application stack or in that tech stack.

[00:04:01] So AI is one of those layers. There's a lot of talk about AI tools or AI products or an AI market, and I think that what we're going to see as companies continue using AI, which is another thing we can talk about, that this hasn't only been happening

[00:04:18] for the past 18 months, is that the things that they're investing in are these business solutions that start to sound pretty familiar to us, right? ERP, CRM, business productivity, HR, finance, and then all of those things are going to have AI under the covers driving new features.

[00:04:35] And so there definitely are things that you have to know about AI to be conversant and to be able to use these things in a masterful way, but you're not going to be selling AI to clients. Clients aren't going to be implementing AI

[00:04:51] outside of implementations that they're already doing. It's shocking. It's not just like a skew that we can be on. And this is where it gets really kind of, I think, the most interesting but also the most difficult, right? Because there's a lot of providers out there

[00:05:05] that just want to sell products. Like that really is the way that they want to approach this. That's a perfectly acceptable business, right? And I don't want to diss on that, but if I'm thinking about this from a high-value engagement, like that's not high value.

[00:05:19] I'm not bringing any... The best I can do is I can make product recommendations and like a product concierge, that's a thing, that's a space. But I think that there's more value when it's something complicated that the customer needs help with. And so what I actually like...

[00:05:39] And I want to sort of start with your sense as somebody who's an analyst in this space. I think that the actual opportunity is like right now. Like I think that there is an actual need right this time for something. And actually like Robbie

[00:05:58] is literally saying the thing that I'm thinking here. So he's saying like, our policies can't keep up with how soon people want these products. I understand their desire to try them out, but we're barely trained and work on themselves. Like that's the need.

[00:06:10] Like literally right there is the opportunity that I think the actual thing to do right now is get your people up to speed and you should be working on two basic things with your customer. The first is like a framework of what your AI is,

[00:06:28] like what your ethics are around them, how it should be used, what your guidelines are. And then the second part is data and the data management. Like that's where my head's at. I think that's now. Do you think I'm right? Am I wrong?

[00:06:40] Where do you think it's going? I mean, I think that's right. The other thing that I would add to that is just kind of a, I don't know, consultation or education or being able to draw out from the customer or the client

[00:06:51] what exactly it is that they want. So if your client is saying, holy cow, I've been playing around with chat GPT and it's amazing and I need something like that in my business. You need to be able to draw out of them what is something like that?

[00:07:04] Is it a chat GPT? I mean, do you actually need a chat bot for customer service or whatever? Maybe. And that's fine. And then you can go down that path. But if you're talking about having that more deeply integrated into workflow,

[00:07:19] it's going to start to get really complicated really quickly. And the customer requirements are, they're going to have to get a little bit more detailed about that and define it more clearly. Instead of just saying, I want the new whiz-bag thing. And so I think that that ability

[00:07:35] to have that conversation and then depending on what they say to kind of lead into a lot of the other parts that you mentioned. So even if they do say, yep, I need a chat bot. Okay, great. So we need data to train it.

[00:07:48] We need to be aware of some of the risks, not just cybersecurity privacy, but also workflow related because these things are based on probability and part of the wonder of them is that they're based on probability. So as much as some of that gets better

[00:08:02] and the hallucinations go away, there's always going to be that possibility there because of the way they work. So what are the guardrails that you're putting in place? How do you have your employees validating that output and then acting on it?

[00:08:17] So there's all those things that go into it. And I think that's kind of what you were driving at. Yeah, a little bit. And I'm going to start asking, like, I don't actually think anybody wants a chat bot. Like, I think the interesting thing about a chat bot

[00:08:29] is that it showed us what's possible. It makes a really amazing demo, right? Like when I want to show somebody like what I can do with AI, like I've got a story like it was outside my... the guy cuts my hair and he was introduced like...

[00:08:42] My barber introduced me to his next client saying, oh, you guys should talk. And he was like, and we were talking about it and I said, well, have you used this? And I actually said, well, what's your industry? And I just asked the question of like,

[00:08:52] what are five great blog posts in... And he was in Corvette management, like the cars and like the maintenance and stuff. And I typed that in and up come nine different blog ideas. And he looked and he said, oh yeah,

[00:09:03] there's like two or three in there that are really good. Now, it makes an amazing demo, but that isn't necessarily how people work. What I'm much more interested in is, is I think Copilot was good in their example, but I think what Apple did this week was great

[00:09:18] in terms of showing the... No, just highlight a section, proofread, change tone, like expand, control, like where it's much more integrated into it. That's the bit. And like, Robbie again with the like, it's deeply getting integrated into this. By year's end, both Windows

[00:09:43] and all of the Apple families of OSs are going to have it integrated into the OS at some level. Like we're going to be dealing with an interactive portion that works with it. That feels useful, right? And I also think that the other thing

[00:10:00] that I'm seeing really useful versions of is data analysis. Like are there areas that you're highlighting that are saying like, yeah, that actually are ones that we're seeing in the market or customer demand for or response survey stuff that says, yeah, this is what customers actually want.

[00:10:14] Yeah, yeah. So, you know, a couple of points off of what you said, you know, when you're talking about this stuff getting integrated and how it's going to change workflow. My simple example for that is like, we've had AI baked into email

[00:10:28] and word processing for a little while where it's doing the auto complete, right? And you're typing and then it suggests an auto complete. And if you want that, then you can hit tab and it finishes it out and then you can keep typing.

[00:10:39] You know, how many of us have picked up that new way of looking at what it suggests, doing that evaluation, hitting tab. And yet then we can get the email done, you know, 27% faster, but there's a learning curve in order to get there.

[00:10:53] And that's just a small example. You know, if you build that out into someone's entire workflow for, you know, processing expense reports or, you know, doing HR or whatever it might be, it gets a lot more complicated a lot, you know, very quickly.

[00:11:07] To your point then about the data stuff, we do see the top request for like AI enabled tools to be data analysis. You know, the top thing that people want to do with AI is data analysis. And that's great. It can probably accelerate a lot

[00:11:25] of what they're trying to do. It can possibly fill in some skills gaps since there are so many people have difficulty finding a lot of those skills. But we've been talking about data and data analytics for quite a while now, dating back maybe to kind of the launch

[00:11:40] of the buzzword of big data. And ever since that time, we have seen that companies really aren't ready to do some of these advanced things that they want to do because they haven't done some data fundamentals like pulling all their data into a comprehensive location, you know,

[00:11:56] or doing some relatively basic database administration. And now with AI, that demand just ramps up where there's a lot of data collection, data cleaning, data input that you have to do for the training and I think so many companies are a little behind the eight ball,

[00:12:16] you know, when it comes to their data practices and being able to really take that next step in preparing an input into the thing they want to do. Now, I'm going to actually spin a little bit because we always talk about people

[00:12:30] like behind the eight ball or like in a race. And I want to actually say, like, I don't actually believe in that as much as I used to on this. It's more, there are people that are well positioned and there are those that have work to do.

[00:12:46] But when we imply a race, we imply that there's like a cutoff or bad results if you don't do it fast enough. And I think what the difference is is that there's just a really, really long period of adoption. You know, we've been,

[00:13:02] there's a lot of technology, right? Cloud technology, right? We've been at this 15 years. We've been having conversations about cloud and there's still people doing cloud migrations that are getting really good value out of their cloud migration, right? Like it's the, they're investing in their business.

[00:13:18] And I think this is one of those spaces where that will be the case as well, that there are companies that are just going to putter along and they're going to do fine. And, but you can do better than fine. And you can actually find ways

[00:13:31] to do more with less, which by the way, considering the labor market and the staffing things may be a thing you want to do. And I look at this and say, like, I think it's more just the, hey, how much do you want to present

[00:13:42] this opportunity to customers versus implying that there's a bad point to this? Am I missing something? Like, is there something that I might be missing out on not moving fast enough here? I think that's fair. You know, we tend to present a lot of things as

[00:14:00] if you're not doing this then you're dead already. And that's almost never the case. I think maybe the kernel of truth that lies in being behind the eight ball or being in a race or whatever is this notion of if you've got in your head,

[00:14:15] if you're a business executive at a client or whatever and you want to do this really advanced data analytics, maybe that's like step 10 and you're on step two, right? And so you're going to have to go through steps three, four, five, six, seven, eight, nine.

[00:14:29] You can't just jump to step 10. And I think that's the value where a provider can come in and kind of help lay that out and say, hey, we can get you here but this is the pathway and there's not going to be an easy button.

[00:14:40] There's not going to be a shortcut. If you want to get to this stage, this is what it will take. This is what it will look like. This is the timeline. This is the investment, you know, whatever. And so I think there has to be some understanding

[00:14:51] of those prerequisites but it's not like you're dead in the water. Okay, so I'm going to put my sort of developer hat on, right? And I literally, I was back at my alma mater recently and talking with people in the computer science department who were kind of laughing

[00:15:07] about the old school way of everything being the waterfall development method, where it was all this very sequential thinking. And if I lean on my software colleagues a lot, they're always agile is all the thing where you have an overriding plan but it's a lot of iterations.

[00:15:23] Isn't this something also where you can kind of move quickly and iterate and improve versus, you know, you said like, well, you've got to get to step 10 but you're only on step two. Like, can't you be more agile and adjustable here? Like, what would that look like?

[00:15:43] Yeah, I do think that there's like a foundation that has to get built before you would start iterating. And I think that what we've seen in our research is that a lot of companies are further away from having that foundation built than maybe they would imagine.

[00:16:02] They kind of know, well, we've got data in the company so we can just start using it. And there's more to do there. And it's interesting that you bring up an agile methodology or DevOps. That's kind of another thing like cloud computing that we've been talking about

[00:16:19] for a long time, but we still see companies trying to refine that approach and trying to improve it. And I think that that probably is maybe the common thread here in what we're talking about. Like, none of this is going to be a single project, right?

[00:16:35] You and I have had debates over the years about digital transformation and whether that's even a thing or not. We recently talked about this on one of our podcasts. I think I've come a little bit more towards your way of seeing things that digital transformation,

[00:16:49] it's not like if you're a company you would say, we are doing a digital transformation project or we're even going through a digital transformation initiative. I think it is a mindset of having a strategic view of technology, not thinking about IT as a cost center anymore,

[00:17:07] a lot of those things. But there is this view of like, this is just the way that we're doing things now. And once we've put all the pieces in place and once we have the right teams, the right skills, the right tools,

[00:17:23] now we are in this constant mode of operation and it's not an overnight thing. It's not a single project with a beginning and an end. We have to figure out how to operationalize this. Is that kind of what you're saying? Yeah, it really is.

[00:17:36] And in a way like it's, Robbie brought it up in a way with his question, right? Because he asked, how do I feel about the integration of the OS? It's like, well, I think the world of technology is continually changing. And it's tempting to always think

[00:17:53] you have to move at the fastest speed possible. And instead, what I'm just looking at saying is, no, I just want you to be moving at a speed, at a non-zero speed of adapting the organization. And so like, from my perspective,

[00:18:09] it's like I can see on the horizon that in just a few months, these kinds of technologies are going to be rolled out at the OS level, right? Like I can say at the Windows level, at the iOS level, at the Mac OS level,

[00:18:26] and on Google and Android and Chromebooks. You can just see all of them are going to have that. And I have to just be working to be ready in a perfect world. Yeah, I've timed that out perfectly and I'm anticipating that and I'm ready with my customers.

[00:18:42] But it's more the, you just have to be moving, right? Because while at one point, you'll think, oh, the users are going to experiment and adapt and go to it. I also know users that things will pop up and they will ignore it, right?

[00:18:57] Like, unless that they actually know the value or have been trained on it, or you've engaged them with the intelligence to make them ready, like it won't prove useful. To your point Seth, if people have not done good work to get the data ready

[00:19:12] to use in an environment, like they may start working with Copilot or with, you know, or Apple Intelligence or whatever they're using with. And it's not that helpful. Like it doesn't necessarily have the right context because the data isn't ready. But that doesn't mean that they're not trying.

[00:19:27] And if you're engaged just continually, you will move with them through, you know, the transformation. Right. Right? Maybe one example that I could offer that I think ties together a lot of what we're trying to say here is going back to that emerging technology conversation.

[00:19:45] One thing that we've been saying for at least five years now is that for an organization who's wanting to accelerate their technology adoption and deal with emerging technologies, one thing that they should have within the organization is a technology evaluation process or team. And if an organization

[00:20:07] had a technology evaluation team already standing up, then AI would have just fed into that process. Right? It wouldn't be like that AI launched, you know, ChatGPT launched and you as a company said, okay, we got to get a tiger team together here

[00:20:21] and start figuring out what AI means. Right? It would have been a process that already exists and it would have fed into that. And the next thing is just going to feed into that too. Right? So there shouldn't necessarily be a thing that companies are doing for AI.

[00:20:36] It's a thing that should have been kind of getting built up to this point and will continue to get built. Right? And I think there are a lot of other examples like that, but I think that's one of the earliest ones in this whole process of understanding

[00:20:50] what technology is coming, how it's getting integrated into the tools that you're using and how eventually to extract the value out of that integration. It's interesting you bring that up because in theory, managed services providers in particular would be really good at process and would be very good

[00:21:10] to be iteratively using process to evaluate that. But admittedly, there's another tension in managed services which is the standardization one. Right? Which is everything must be standardized, everything must be packaged, everything must be a monthly recurring revenue, bit only. Like when you get into that bit,

[00:21:27] you're not actually motivated to use process to change things because that breaks things. Right? It's an interesting tension and it's why at times I look at providers and I say, well, are they actually going to be the ones to do this really well? And why my message

[00:21:45] in particularly lately has been around the, remember consulting is not a bad thing. I know it's not monthly recurring revenue purely, but that's the bit where new chaos gets introduced because you've talked to your customers, you're billing them in a different way

[00:22:03] that allows this to be more free form and then leverage your process stuff. So let me pivot a little bit on that process stuff because I want to get your take on the importance of governance. I've been talking a lot on the show

[00:22:18] about the idea of AR frameworks and putting frameworks together. Like how important and where does it fit in governance into this space? Yeah. It's interesting that you bring that up because our outlook report from last year where we kind of say, here's our top 10 big trends

[00:22:33] that we're expecting over the coming year in the fields of cybersecurity and data management, we highlighted governance. And I think that when you put governance and cybersecurity side by side, everyone's mind immediately goes to regulatory compliance, right? You know, your HIPAA, your PCI, you know, increasingly more

[00:22:54] and more regulations around doing digital business. And that is one part of it. But the meaning that we had with governance, and I think the meaning that you're getting to is it goes beyond that into these are the best practices

[00:23:09] and this is what we are going to follow. So we're going to say what we're going to do, we're going to do it and then we're going to measure it, right? And if you don't have all of that in place, then you can't measure the output.

[00:23:20] And as companies are trying to move to a more strategic view and they're trying to get away maybe from some of that cost center mentality, then they're having more questions around if I make an investment in technology, what do I get out of that? And in some cases,

[00:23:36] it can't always be a standard classical ROI calculation, you know, especially in an area like cybersecurity. But in some cases, it can get a little bit closer to that. But I think that's a very new muscle to exercise for a lot of even, you know, CIOs,

[00:23:51] internal IT departments, and then certainly, you know, managed service providers and other third parties that are kind of participating in the technology strategy with their clients. You know, it's interesting because I think it's all of that. And I think it's one additional thing.

[00:24:05] And that one additional thing is, is I actually think you want to put together, particularly around AI, a set of guiding principles that actually help your staff and your team know how to not get themselves into trouble. Like, you know, one of the things that like,

[00:24:23] I'll use my, you know, little media organization as an example. So I have an AI statement on my, in my ethics listing of the, like the way I use AI, that it is like, I don't mind saying, AI helps me with a whole bunch of stuff.

[00:24:36] But the human is me. I am responsible for everything that's published. Nothing goes out without me at least looking at, you know, looking at it, reviewing it, editing it, whatever that process is, that that is a guiding principle for my company, MSP Radio,

[00:24:51] my shows, The Business of Tech, is that, yeah, I'm totally using all these tools. You should expect that there's a whole bunch of them behind there. But a human looked at it and that's a guiding principle. Apple just puts out theirs, right? Like the fact that they,

[00:25:03] they want to respect privacy and that they're going to use, you know, they're going to test for the negative cases and the bad outcomes as part of their guiding process. Would, you know, is this an area where you're seeing in, as you talk to

[00:25:19] and look at these MSPs, are you seeing them have these kinds of discussions or is this, you know, super breaking kind of, you know, early, too early to say? It's kind of transparency, right? I mean, I think that that's what you're talking about. And that's a term

[00:25:33] that we've used a lot when it comes to data usage. You know, this is how we are using your data. And I think to be able to make that kind of statement kind of requires a degree of acumen that many people might not have built.

[00:25:49] You know, whether you're an MSP or a client, and in defense of MSPs, I will say a lot of clients probably aren't at the point of fully understanding what they even want, right? And so then to ask an MSP to come in could be, you know,

[00:26:07] kind of challenging. And I think this is where, you know, we've really been encouraging MSPs to be thinking about their value. And I think that plays out in marketing activities. It can play out in consultation opportunities like we've been talking about. But to kind of know

[00:26:22] what that value is and to be able to understand all of the different puts and takes that are happening with technology these days and to talk intelligently about the trade-offs, to talk about the different things that you have to watch out for in putting together a statement

[00:26:37] like we're talking about. I think there's so much opportunity there. And yeah, it may be a little bit of an uphill climb because so many clients, I think, are still in the midst of that transition from tactical IT to strategic IT. And even if they're,

[00:26:54] you know, maybe the halfway point on their way to strategic IT, maybe they view the MSP as still one of those tactical pieces. And again, to us, it gets back to understanding your value and being able to describe your value and describe that to the client.

[00:27:12] And if the client still isn't buying what you're selling, then, you know, you can keep trying. But I think there is a lot of work to do there in that understanding of value and being able to communicate it. Well, Seth, I mean, I almost think that

[00:27:24] you've literally just outlined what the key value is and what providers should be working on. So I couldn't think of a better place to sort of wrap our thinking. You've just outlined the bit that people should be working on now with their customers.

[00:27:39] So really appreciate you joining me. People can find you. What's the name of your show? And tell me a little bit about it for listeners that may not be aware. Yeah, so our podcast is Volley. A lot of our conversations sound a lot like the one

[00:27:51] that we've just been having for the past several minutes. It's me and my colleague, Carolyn April. We co-host. We have guests on. And we just talk through the technology topics of the day. A lot of times, we will feature our own research.

[00:28:05] So that's one place to find us. And then you can find a lot of the research on CompTIA's website, comptia.org. Well, I am a subscriber. Seth, thanks for joining me today. Thanks, Dave. So I want to talk one thing about an interview

[00:28:17] that's coming up this coming weekend. Petra Molinar, who's a lawyer and anthropologist specializing in border technologies, joined me on the show to discuss the impact of AI-driven technology on migration, particularly in her book, The Walls Have Eyes. We chatted about the ethical implications

[00:28:35] of using technology at the border and the importance of considering those real-world impacts when innovating, just like Seth and I were talking about. Here's a preview of that interview. Six years now, trying to understand how new technologies are playing out at the border and precisely how big

[00:28:52] of a business this really is because we're talking about a billion-dollar global industrial complex that is implicated in how technologies are developed and deployed. And what I find helpful sometimes is looking at it kind of from a temporal perspective, trying to understand what happens when a person

[00:29:11] is crossing borders because at this point, your entire journey is impacted by tech. Stuff happens before you even move, things like social media scraping, other types of data analytics. There's all sorts of things that are happening right at the border when it comes to surveillance,

[00:29:27] really experimental AI projects like robodogs or lie detectors and things like this, things like voice reporting, different types of carceral technologies and immigration detention, and even some of the less sexy stuff like algorithmic decision-making that happens under the hood that a lot of governments are experimenting with.

[00:29:46] You'll find this interview fascinating. We talk about a whole area of technology I wasn't aware of, and we're always looking for interesting use cases to understand how things might be applied or what happens when the laws are a little bit more gray. My supporters already have this interview.

[00:30:02] It's dropping for the public on the weekend, but if you want to listen now, you can sign up as a Patreon at msbradio.com slash, or sorry, patreon.com slash msbradio, and you can sign up for it to get that early. If you're interested,

[00:30:15] I really do encourage you to listen to that. I want to thank SalesBuilder 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 then there's you too

[00:30:33] can get your name mentioned on the live show. It's a simple monthly subscription. Visit patreon.com slash msbradio to learn more. Listeners, if you've enjoyed this, you can support the show in the best way possible. Like, share and follow on your favorite platforms. It really makes a difference.

[00:30:49] Tell one friend you learned something interesting today on the business of tech and have them sign up or support directly on Patreon with our give what you want model. You set when you think the content is worth. If you're listening to the recording and you have a question,

[00:31:02] send it in at question at msbradio.com. We do answer listener questions. We'll be off next week from the live show and back the week after. Look forward to those conversations and answering your questions. Thanks for joining me for the business of tech lounge

[00:31:17] and I will see you next time. .