Building Citizen Services for the Age of Agentic AI with Insight Public Sector

[00:00:00] Do you need AI agents that you can trust? Well, with an AI data layer providing real-time connection within your data platforms, you can trust your agents to provide accurate solutions. So, scale your business by trusting your agentic AI accurately getting the work done for you. Trust its capabilities with Denodo. And you can do that by simply visiting denodo.com to learn more.

[00:00:26] What happens when AI stops just being another pilot project and starts becoming a part of the infrastructure? Infrastructure that powers public services. Today, I'm going to be talking to the public sector CTO at Insight Enterprises.

[00:00:46] And alongside leading technology strategy for public sector organisations, my guest also teaches graduate-level cybersecurity cloud computing and AI programmes. And I think all of these things give him a unique perspective on how emerging technologies are both reshaping government operations and the workforce of tomorrow.

[00:01:10] So, today, I want to move beyond the headlines and examine what AI adoption actually looks like inside public sector organisations. Because my guests will argue that we're entering a period where AI is moving from experimentation into operations, with agentic systems increasingly supporting everything from citizen services and permit requests to internal workflows, knowledge management and service delivery.

[00:01:36] And we'll also discuss why many governments are moving away from large, centralised AI initiatives in favour of smaller, more focused deployments. Ones that can deliver results faster, provide stronger contextual understanding and better align with departmental needs. But along the way, we will also explore how universities are adapting to a generation of students

[00:02:03] who already use AI on a daily basis and why governance must become part of operational processes rather than separate compliance exercises. And overall, how can the public sector balance innovation, trust, security, accountability? We're going to cover a lot today. So, if you've been wondering what practical AI adoption looks like, not in the private sector, but go beyond demos and keynote presentations

[00:02:32] and take a thoughtful look at what it means to the public sector and where innovation is heading and what lessons organisations everywhere can learn from it, I think you're going to enjoy this one. But enough for me. Let me introduce you to my guest right away. Thank you for joining me on the podcast today. Can you tell everyone listening a little about who you are and what you do? Yes, thanks for having me. I really appreciate it. And I am Insight's public sector CTO.

[00:03:02] So, I work for Insight Enterprises. I take care of the Insight public sector technology. And it's an interesting position because what we're seeing here is a huge change as it relates to AI. And we'll talk more about that, I'm sure, as we get into the podcast. But in addition to being a CTO for this organisation, I'm also an educator. So, I teach at Northeastern University in the Boston area. So, I teach graduate school in cybersecurity, AI, cloud computing.

[00:03:30] And I also am an academic director for Wake Forest University, where we've created from the ground up an AI strategy and innovation program in the educational space for a professional studies program like extension school model. Again, for master's degree program students that are looking to reinvent their careers and professional journey. Or for students that are just looking to get more knowledge. And it's kind of interesting. So, I'd love to talk about that some today too.

[00:04:00] Yeah, it's a real interesting space right now. And I'm glad you're joining me because I've been to 15 tech conferences this year. And predictably, everything's about agentic or agentic AI. And by that, I could name companies from Adobe to Zendesk, all saying the same thing. And you're predicting the agentic operations will actually move from pilots into mission critical infrastructure now. So, what does this look like in the public sector environment?

[00:04:27] Because I get to hear a lot about the private sector and how they seem to be racing ahead. But tell me about the public sector and how different will citizen services begin to feel as a result? Yeah, so in the public sector, and we've seen this probably globally, right? It's not just within the US, but I think across the globe, we've seen cutbacks in terms of resources and funding for public sector style programs. Even federal government programs follow the same model.

[00:04:54] But with that, we are almost reliant on the fact that we need some kind of automation, some kind of intelligence, some ability to be able to do more with less. And it's not just training, it's leveraging the AI capability. So, I think AI came at the right time, or at least our current version of AI, at least how we describe it. Because I can get into a long diatribe about how long AI has actually been around, but that's okay. Most people really just treat it as the last two years.

[00:05:22] But in the grand scheme, we've moved from sort of trial of what can AI do for us, like with a chatbot, to can we create systemic kinds of workflow processes that will allow us to be able to do more? And it's really resonating in the public sector. So, I think they're more inclined to look at what can these services do for them and how can they treat them as if they're standard workflow processes within their organization.

[00:05:47] Just yesterday, for example, I was in a Microsoft technology center talking about these specific kinds of things. Like, there's a bit of a trust issue in terms of should we let AI do our work for us? But I think there isn't, and I don't want to sound too pedantic about it, but there is a bit of a, you have no choice. You kind of have to move down the path of we must serve our citizens. And so, how are we going to do that most effectively? And it's that transition to operations, if you will.

[00:06:15] And as AI gets better, as it starts to become more integrated into the platform stacks like, you know, the Azure Stack or Google Stack, et cetera, we're starting to see better and better capabilities. And I think people are looking at it more from a trust perspective. And trial and error is going to get us there. So, a lot of organizations will go through this maturity model phasing. And I do think in this year it's starting, and I think next year it's going to get even more prolific within public sector.

[00:06:42] So, people will start relying on AI services to help them with human in the loop, of course. They'll put that in there. But that's really, I think, the way things will go over the course of the next year or so. Yeah, I completely agree with you. And for people listening thinking, yes, there has been a lot of talk about chatbots over the last few years. I think it's important to highlight what you're describing here is something much deeper and bigger than chatbots.

[00:07:06] So, what separates a basic conventional interface from a true operational AI agent that's embedded into government workflows? Because people listening probably got a few ideas or bad experiences with chatbots. But just to help them understand what we're talking about here. Yeah, I think the, and I usually treat it as normally chatbots are really targeted toward the individual. Or, you know, maybe it's even for the worker and workforce productivity. So, it's an individual usually interacting to try to learn something. Or even a citizen, right, connecting to.

[00:07:36] We've done several of these over time where, you know, I want to understand a little bit more about how do I get a permit within a particular county. And these chatbots can help. But as you said, sometimes it's really frustrating because maybe it's not giving me the answers that I wanted. Or worse, it could be, you know, there are 27 languages spoken within my city or town or state. And I can't actually type in English very well, so I can't get the answer I want kind of thing. So, there's those kinds of struggles, I think, that come with it.

[00:08:06] So, that part, I think, can yield a bit of, if you don't really think about it very clearly, it can yield a mediocre response. When we think about workflow and workflow processes, I think what's really important there is the agentic workflow model or when we're trying to create more enhanced operational processes, a well-thought-out process can actually help us because we know the expected outcome. And the thing that I do like about the agentic world is their goal-seeking.

[00:08:33] So, it basically will work until it achieves the goal to an extent with human-in-the-loop models. So, you can basically ensure that you can interact. I'll just give you an example.

[00:08:42] Like, Austin City, they just published a new – it's not really a law, but it's basically – call it a statute that says that all decisions being made for a citizen or for a citizen to consume must be approved, if you will, or generally by a human. So, it's like we're not quite at the level where we're going to let the agentic AI or operational processes do everything.

[00:09:11] But I think what we can do is we can define processes really well. We can leverage some of this automation through agentic environments. And then we can test it and iterate and sort of release more and more control once we develop trust and all the best practices that we typically use. So, I think we need – we're beyond the it's a cool toy, let's stop playing with it, to let's make – let's get serious about it and productionalize what it is.

[00:09:36] And I think that's really the big difference between sort of a chatbot that's really nice to, you know, interface with and get answers from versus an operational process flow that can really produce business impact and efficiency. And higher education institutions, that's an area that's also under incredible pressure right now to modernize our operations and student services flow, etc.

[00:10:00] So, I'm curious, from what you're seeing and hearing here, how are colleges and universities beginning to apply AI? And also, what lessons are they learning that maybe governments or people in the public sector should be paying attention to? Yeah, exactly. It's a great question. And because I have direct exposure here, I can give you my perspective on it and from maybe two different vantage points. But the first one is usually – I think people fall into two particular camps.

[00:10:29] There is the groups of educators that are more focused on we don't trust AI, we don't want to bring it into the classroom, we want to make sure people are not using it, and it's something where it's thought of as a mechanism for not – I'll call it not real understanding. You know, so they feel that it's something that equates to cheating. I hate to use that word, but they sort of see it in that way.

[00:10:56] And they believe it prevents, you know, true education. The other camp is – embraces it. And so this is ultimately what my Wake Forest program is. So I have, I think, eight or nine faculty that are working with me to create courses that are focused on helping people understand AI strategy and innovation. And in that realm, we embrace AI. So we're basically saying you must use AI or you're going to use AI to understand how to enable your understanding and capability.

[00:11:25] And I think some of what we'll probably talk about later is really focused on how can you go faster? How can you experiment? How can you really understand? Because the art of understanding and learning is really about experimentation, exploration, innovation. That's really what I think we need to do. So there's really two big camps in that area. Yeah.

[00:11:43] And as a CTO for a public sector organization, I can look at it from the perspective of a slightly different vantage point, which is the other piece of this is with the advances in the technology. So we're close to all of the big partners of the world, the Anthropics and the Open AIs and the Microsofts and the AWSs and the Googles. And you look at it from the perspective of, well, what capabilities are there?

[00:12:08] What can – how can you align it to the objectives and the goals for each one of the students? Because we do work with the education sector as well. And so you're really just sort of helping to enable them to understand what needs to be done independent of their pedagogical viewpoint on what it is that they should be doing. So I think that's – it's really this interesting little mix of how to treat the – I'll call it the disruption that has occurred clearly.

[00:12:34] And to your point, Neil, is the fact that within the education sector, I think they're pretty far behind. So students have exposure to more advanced technologies than are being made available through the educational institutions. And so I think that has to start to change more rapidly. And part of its culture, part of its technology adoption, and part of it is really understanding the – how do they want to create this pedagogical environment, which allows you to be able to embrace AI, which I think we have to do.

[00:13:03] I don't see any way around it. I also think it would be almost criminal, especially when preparing students for a new generation of work, job roles that don't even exist yet, to almost pretend that AI isn't there and not happening and we're not going to do that. Because you can't swim against the tide on this stuff, can you? No, yeah. That's a really great point because it is that. It is ubiquitous or it will be ubiquitous if it isn't already.

[00:13:29] And honestly, even five years ago, 10 years ago, even AI was sort of making its way into our phones, into our web applications, and it was being embedded as machine learning. But we really didn't think about it too much. Even things like GPS systems and mapping and shortest path algorithms and optimization, that's all kind of intelligence that's built on algorithmic style machine learning.

[00:13:57] So we've been using this for many, many years. I think people just didn't realize it. But yeah, you're right. It is part of society and we can't avoid it. And the jobs of the future will require you to be able to do it. And I'll give you a really interesting example of how this works. Like one of my courses, what I typically do is, and it depends on your pedagogical view. And I think people listening to this might, you know, they can feel free to call me if they want to talk more about it.

[00:14:23] But my perspective is, if I'm having them use AI and they have to use it in the real workforce, and I give them an assignment that is, you know, don't write a paper because I know that that's silly. But if I give you, I expect you to create maybe a solution or a project that's going to be able to use certain kinds of technology and you're using vibe coding or whatever it might be. And it would normally take someone, say, 30 hours to complete or 40 hours to complete. Well, you can't do that in the timeframe that I'm giving you in the course. So you have to use AI.

[00:14:52] So I'm sort of forcing them to understand how to use AI to accomplish more. And if you don't have the critical thinking in place to know what to do next and how to navigate yourself through the process, then you won't succeed. So I'm sort of taking that approach, which is really good experiential knowledge, right, for when you get into the workforce. And it's, you know, it's early days. I mean, I'm hoping that it works. Several of my students have told me that that really works well for them and helps to prepare them.

[00:15:20] But, you know, hopefully it turns out to be the wave of the future. Who knows? Yeah. That'll be interesting to watch how things evolve there. And you've also mentioned the rise of agentic front doors for public services. So how do you ensure that systems like that can improve accessibility and efficiency without creating even more frustration or essentially losing that human touch that citizens still expect to be at the heart of everything? Yeah.

[00:15:46] And that's a really interesting one because it's like, I think initially when I thought about it, I was like, wow, it can create this standardized, consistent approach and people can get what they want and it will understand, you know, the intent of what they're asking. And it can do all those things, which is really great. But unfortunately, and I just saw a report, I can't remember who wrote the report, but as agentic AI sort of makes its way into e-commerce, people don't trust it.

[00:16:14] So there's only six, about 6% usage because people are sort of don't trust the fact that AI is enabling or helping them. And I fear that that still might happen in the front door. So if I know an AI is going to be responding to my query, my question or whatever, you know, interfacing with me as a citizen, I might push back.

[00:16:35] So I think there is a little bit of work that we have to do in the public sector to be able to ensure that we don't just create trust and confidence within our own public sector or workforce, but we make sure citizens understand that as well. So it's sort of, it's sort of educating the citizenry in terms of we're providing better service to you and showing them how, and, you know, you can do that with public service messages and things like that. But I do think that there's a cultural part of it that has to come with the technology part of it.

[00:17:04] But I do think that it can create huge impact because I was talking about this meeting with a city in Massachusetts yesterday and talking through particular use cases. We probably came up with 10 different use cases, which were how can we create a better citizen experience? And all of them were really just related to autonomy. So customer or the citizen calls in is able to get what they want. They want to be able to like, how do I, how do I know whether I can put a fence in my yard?

[00:17:32] Is it too close to an easement or too close to my neighbor's yard, whatever. What are the rules? Even simple things like that, if you can trust it, can change the experience for the citizen. And one of the most interesting shifts that I've read that you highlighted is this move away from large agency-wide platforms towards smaller, smaller, modular AI pods of sorts. So I've got to ask, why do you think monolithic approaches have struggled and what makes these smaller deployments more effective?

[00:18:02] What are you seeing there? Yeah. So I'll answer it in two ways. And the first one is contextualization because I think the large monolithic platforms tend to look at it. You can get a good 30,000 foot answer, if you will. But if you're really talking about something very specific to the local town or government organization, that is very specific to their laws, statutes and policies. And that contextualization, I think, is important.

[00:18:31] So do I really need to know in terms of, and I'll even treat it as a large language model. Do I really need to know how to whatever, design a pharmaceutical drug? Does that need to be in the model at the time at which I'm trying to answer whether or not you can exercise a specific policy within the government organization? Probably not. So sometimes it's overkill. So I think these more specialized environments can really help.

[00:18:55] And token costs will start to become a thing because cost is all important to us, especially within state and local government or even federal government because of the fact that there's limited budgets, et cetera. Nobody wants to pay more taxes for AI, I bet. And then the second one is really related to, and this gets to serving the public.

[00:19:16] I'm going to go to IoT and the whole world of extending the reach of AI all the way to where the citizens interact with it. So how can I provide better security in a smart city to be able to protect my citizenry more effectively through AI or intelligence directly in cameras? And I can have agentic workflows running directly in the camera or a near-edge solution.

[00:19:44] So I can really see these sort of more subject area, contextual, specific kinds of AI that are workflow-oriented agentic deployments that allow us to be able to provide better service overall. So, and I think it's a good distribution model generally too.

[00:20:02] I mean, I love the fact that we have all of these huge models that can do all these wonderful things, but sometimes it's just about solving the problem that we're trying to provide impact for our particular local government or town or public sector entity.

[00:20:18] And from your experience and your perspective right now at Insight Enterprises, how are public sector organizations balancing speed of deployment with strict compliance requirements and everything else that they operate under where things can be a little bit more difficult? And there is that reputation that they have to move much slower than the private sector. What are you seeing here? Is that a myth? Are things improving? Things are definitely improving. And I think it's because of what we were talking about earlier.

[00:20:44] So I'll kind of take the last part of your question first, but it's really, you can't really avoid the fact that this is a societal disruption. So even in the public sector, they can't just say, well, we'll wait three years and then we'll finally figure it out. You can't. People are using it. So even if you are not providing, say, these services internal to your workforce, then someone else is actually, they're going to use some other AI technique and they're probably going to be doing things like data loss.

[00:21:12] You know, they're going to using that data within whatever they can find for free to be able to help them do the job more effectively. So that's a, that can be a big problem. So the way to combat that one is to make sure that you provide these services. So give them the co-pilot licenses or give them the chatbots, let them create a gentic environments. The second part I'll address is really just related to how is it that we really understand how to create these kinds of environments and workflows. And so training and enablement, I think is a big part of this.

[00:21:41] And it, it is really this sort of culture of innovation and iteration. So one of the things that we do with a lot of our customers is, you know, it's not a monolithic project in terms of like, let's just say that I wanted to create better visibility into, are you familiar with 311? So 311 is basically the number you call when you want to report some kind of incident in the, in the city. Not, not like a 911 incident, but something that's like, you know, oh, there's a pothole on this road or, you know, there's trash bins in the middle of the street.

[00:22:11] But there are 311s are pretty important for many cities. And so they want to be able to do things like automate the, you know, reception of this, make sure that you can create out a form, put it into the right system, et cetera, et cetera. So efficiency there is really important. And they, sometimes they can't even get to all of their calls than the 311. So this allows them to be able to become more efficient. And the ways to really deal with it is sort of an agile model.

[00:22:36] So even though we're writing AI techniques, we can do sort of incremental buildouts of agentic. You can start with even simple chatbots. How does it work? And then move to the agentic model, simple agentic, and then go more sophisticated agentic models to operationalize it.

[00:22:53] So I think that iterative pattern is one way to sort of get us from conception all the way through to production without necessarily, you know, saying we have to commission a nine-month or a two-year project in order to be able to get there. So I think that's another thing that helps us to move faster within these environments. You know, just get going. And I tell people this all the time. Maturity model-wise, it's like, just start, try it.

[00:23:18] And, you know, there are many examples of this where people are just experimenting, and I think that's really the right way to do it. And then the last part is to just culturally within any public sector organization or even within any industry organization, it's really looking at AI as really just sort of the next generation of, I'll call it application development or solutions development. Because it's not that far different. Yes, it's using convolutional neural networks and all this other cool stuff.

[00:23:46] But it's really just another way to exercise business logic. And so you already have the data, you already have the network, you already have the security in place, you already have some infrastructure. So you're just adding on to what you already know. So some people think, oh, we have to start from scratch. No, you don't. You're kind of sort of picking up, you know, on the things that you already have done and then moving forward. So it's sort of an incremental value add. It's not necessarily a start from scratch model. So I think that's an important point for many organizations to realize as well.

[00:24:16] And traditionally, governance has been treated as a policy layer that sits outside of implementation. And refreshingly, I was reading that you've suggested it actually becomes embedded directly into operations. So for people listening, what does operationalized security look like on a day to day for teams that are building and deploying AI? Because I think it's such a great approach. Yeah, yeah, yeah. Yeah. And I forgot to answer that part of your last question.

[00:24:42] So I'll start with the fact that this is the laggard in all of this, at least in the US. And it's probably not true in other countries, because I think Europe is doing a better job here. But what's happening is that there aren't a lot of regulations in place. And I think some cities and states are starting to enforce their own or create their own laws, regulations, because we don't really have a great national policy as it relates to use of AI. So we're starting to see some of that.

[00:25:10] So there isn't a lot of policy or laws that are on the books that are going to impact the way that you implement these kinds of solutions. So getting to the operationalization part of it. So, you know, keeping that in mind and knowing that eventually you will have a lot of regulations, because I think at some point the shoe will drop and we will start to have a lot of regulation that comes in place, especially for global companies.

[00:25:32] But I think what we see now is there are specifications, like you look at the NIST AI RMF or even the NIST specifications, which already have ways to sort of treat data and identify what risk might be in the system specific to AI with the AI RMF.

[00:25:48] But it's using these frameworks can help you to operationalize the way that we think about what do we do when specific kinds of vulnerabilities happen or what should we look for or where might we be susceptible or what actions should we take as it relates to things like ethical and responsible use of AI. So I think the NIST has done a really nice job and there are others. CISA is also in this boat.

[00:26:14] And I know Europe has done and the ISO standardization has done a great job in terms of identifying frameworks to create that level of operationalization so that you're not exposed, so that you're not say releasing something that, you know, Neil created and said, oh, I think this is ready for production. And you just released it. And all of a sudden it's, oh, you know, we just exposed, you know, half of our data to the web.

[00:26:36] So it's, I think following those sort of regimented industry best practices or what we call national standards tend to create an operational environment that allows you to adhere to best practice principles and keep things safe. Because I think that's an important part because earlier we were talking about things like how do I get customers or citizens to use these products? It's about trust.

[00:27:00] So if you created a web app, for example, and I created something where I expose citizens data, no one will ever use it again. So once I've lost their trust, I'll never get it back again or it'll take a long time. So I think these kinds of environments really do help. And it's something I teach actually. And as part of my, you know, cloud adoption frameworks and AI strategy, make sure that you leverage the frameworks because you don't have to reinvent it. These things are out here for you to use.

[00:27:28] And you mentioned frameworks like NIST, AI risk management framework there and how they're helping governments think about AI risk. But I've got to ask, are these frameworks keeping pace with how quickly technology is evolving? Because we look in the last 12 months alone. I know last year, nobody was talking about agentic AI tech conferences. This year, it's all everyone's talking about. The pace of change is so quickly, isn't it? It is. Yeah.

[00:27:54] And I do think that there's a little bit more agility and especially in the NIST AI RMF. I know the NIST cybersecurity framework was maybe a little bit behind in terms of the way it worked, but AI RMF is moving faster. And then I even sit on a part of a board for NIST, which talks about education and training. And so there's addendums that come out quite frequently to try to keep up with the evolution within this area. And I'm also seeing some leapfrog.

[00:28:24] So there's basically policy frameworks like CISA just came out with an agentic security model. So I think you see different organizations starting to think about additive security controls and protection frameworks because I think it's going to become a big business ultimately.

[00:28:43] And it's not just about becoming a big business, but I think it's the threat of potential, not existential necessarily, but I mean, you know, it is going to happen at some point where there's going to be a big incident. And someone's going to say, well, why didn't we follow this framework? Because it wasn't available. And so I think we have to make sure that we can get ahead of it as much as we can. But your point is well taken.

[00:29:08] It is definitely there's probably a lag time and it could be three to six months behind the current state of the art. So I think we have to make sure that we are at least following principles, best practice principles. And then we might have to extrapolate for a little while until things get written down as strict policies in terms of implementation. Yeah.

[00:29:28] And another area that has a reputation for being cumbersome and frustrating is, of course, procurement, especially in the public sector, which has traditionally been very slow and risk averse. So how do you see AI changing the way governments evaluate and buy technology, especially when outcomes are so much harder to define up front?

[00:29:48] Yeah, I think on, well, I'll say, unfortunately, it's really difficult for procurement because, you know, as much as I love the tool vendors, they are confusing everybody. So it's like they're trying to look at it from the perspective of we can do all these things and, you know, capabilities versus true solutions. Because I think it's really difficult to know what do I need in order to be able to accomplish my particular goal or my use case.

[00:30:16] So it's like you never really know what it is that you need. And that is probably intentional because I think it's a market share question, you know, where, you know, the vendors are looking to be to land that big project, which is like all of our AI within the state of Massachusetts, for example, is open AI, which I think is a fact. So they basically had this big agreement with open AI. But how did they select open AI? Well, there was probably some kind of analysis that was performed.

[00:30:41] And if you can do apples to apples comparisons, that's really hard because, you know, they intentionally don't allow you to do that. You can't just say go down line by line and say this is what I'm going to do. So I think for procurement, they have to collect as much information as they can, get the technology people involved, maybe see demonstration. So I think their job is really hard. And then I think on top of it is the thing that is not really clear either is you may not know what does use look like.

[00:31:11] So like two years ago when Copilot first came on the scene and people bought it, you know, you'd buy whatever, 5,000 licenses, and maybe only a third of them were ever used because people didn't know what to do with it. So you never know quite what to expect in terms of the overall cost profile. So if I'm in procurement, not only do I pay licensing fees, but I also want to ensure that I have usability. So if I spent all this money, am I getting a good use out of it? There's another big problem on top of it.

[00:31:39] And procurement can't guarantee that, but they still are accountable to the fact that they spent all the money, if you will. So I think it's a hard game right now because of, like you said, the changing technology, the advances in sort of our understanding and learning, and the ability of the organizational culture to adapt to be able to leverage it. So if I'm in procurement, I kind of want to go through this checkboxing and say, do we know how to use it? Do we understand what the benefits are?

[00:32:07] Do we know where the vendor is going in terms of innovation and what the roadmap looks like into the future? And does that align with where I'm going in my three to five-year plan, if I can see out that far, maybe one to two-year plan? But yeah, I think there's a way to do it, but you've got to be pretty good if you're sitting in that seat. And although there are 3,500 miles between us today, I think there are so many similarities in what we're seeing in the public sector here in the UK and the US.

[00:32:35] So looking ahead over the next, what, 12 to 18 months, what are the clearest signals that a government has successfully moved from experimentation into real scalable AI adoption? What are the signals there? What should we be looking out for? I think it really focuses on, and I don't want to make it sound like it's all about sort of executive level adoption,

[00:33:01] but what I've seen in terms of true adoption, if you have states, well, and I'll talk about it in states, you can look at other sort of hierarchical representations of government sector or public sector, but I think when you get executive level buy-in and commitment to be able to provide these types of services, that is really the recipe for success. The technology I don't think is as difficult, but when you get that cultural adoption,

[00:33:27] you can usually then sort of allocate enough resources to be able to drive the adoption. And I'll give you an example. Within state of Texas, they have a mandate for all, and they provide training credits for all state employees to be able to understand how to use the technology to innovate within the state. So this is something that I think is really valuable. And other states don't do that. So you're potentially at a disadvantage.

[00:33:57] So I think it really comes from the top a bit. Of course, you need talented resources. You need the right technology. You need all those other things, but they're necessary but not sufficient. Like if you don't have that organizational support and structure, and I'll call it even allocation of funds, you're not really going to get there. And another thing I like to talk about is visibility. Because I think sometimes when you head down a path like this and you do something, you know, I've automated a permit workflow. Now it's more efficient.

[00:34:25] It saved me whatever, 350 hours a week or whatever. Maybe not that much. But in general, this ends up being something that you have to communicate it. Like you have to recognize it. So who is actually publicizing the fact that we did create efficiency? And it's really just sort of feedback mechanism that allows you to say successes yield more success. So I think it's a bit of a cultural thing, a bit of a technology thing. But I mean, I wouldn't say we're quite a technology parody. But we're kind of there, right?

[00:34:55] I mean, you know, am I going to use a Gemini gem or am I going to use Microsoft co-pilot or whatever, you know, a co-pilot agent? Okay. They're kind of all doing the same thing. And I'm not disparaging any of the vendors at all. I'm just saying that it's not really about the technology. It's more about the ability to be able to recognize the fact that you need to create organizational impact within the state, city, local or town organization.

[00:35:22] And then you have to have the executive level support and resources to do it. And I think that's the important part. Does that help? Does that kind of answer your question? It really does. And I cannot thank you enough for taking the time out to come on here and just talk about the changes that you see coming to AI, especially in the public sector. It's an area that I don't talk about enough. And I'm hoping that we've talked about it in a language everyone can understand and get further by and then a bit of engagement and keep this conversation going.

[00:35:50] And for anyone that would like to keep that conversation going, want to find out more information about everything that you're doing over there at Insight Enterprises too. Where would you like me to point everyone? Yeah. So you can go to ips.insight.com, which is our, I'll call it our public-facing, public sector link to Insight. Or you can find us on LinkedIn at insight-public-sector. Or you can find me on LinkedIn, ctaglianti, if you want to follow what I'm doing.

[00:36:20] And I want to thank you so much, Neil, for having me on today. It's been wonderful to talk about this and a great conversation. And certainly love to talk again if you have the interest to do so. 100%. Thank you for taking the time. There was an open invite for you to come on any time. And we did cover a lot there in 30 minutes from agentic operations, moving from pilots to the back office, how AI governance is becoming operationalized security,

[00:36:45] and departmental-scale compliant AI footprints are outpacing agency-wide platforms. So many big takeaways. If ever you want to come on and drill down a little bit deeper on any of these things or something new, there's an open invite there. I will include links to everything that you've mentioned. And for people listening, please go check those links out. And let's keep this conversation going. But more than anything, just thank you for bringing it all to life today. Appreciate your time. My pleasure. Thanks so much.

[00:37:12] I love my guests' focus on culture rather than technology. Because, yeah, the tech itself, it is advancing at remarkable speed. But again and again, we come back to the questions of trust, adoption, education, governance, leadership. All these human factors will ultimately determine whether AI delivers meaningful outcomes or simply becomes just another underused technology investment. And nobody wants that.

[00:37:42] And I loved his vision of agentic operations becoming part of everyday public service delivery. But the idea that AI can help citizens access information faster, automate routine processes, and support overwhelmed public sector teams. All these things feel more tangible than many futuristic AI discussions that I often get to talk about. And his comments on education particularly struck a chord as well.

[00:38:10] Because universities, governments, and businesses are all wrestling with the same reality. Yep, AI is already here. The challenge is no longer whether people will use it, should they use it. It's about how they learn to use it effectively, responsibly, and productively. So my takeaway is successful AI adoption appears to be much less about building massive platforms and more about honing in on specific problems.

[00:38:40] Start small, focus on outcomes, learn quickly, and build trust through measurable success. It's the belts and braces of the IT world that I grew up in. But as always, love to hear your thoughts on this one. Do you think public sector organisations are moving quickly enough to embrace AI? And what role should governments play in balancing innovation with accountability as these technologies become part of everyday services? TechTalksNetwork.com. Love to hear your thoughts on this.

[00:39:10] But I'm afraid we're out of time. So I will bid you farewell. I will return to your podcast feeds. Same time, same place tomorrow. Hopefully, I'll get to speak with you again then. Bye for now. Bye for now. Bye for now. Bye for now. Bye for now.