EP 73: The AI Tipping Point for Government featuring Jeremy Wilcox

[MUSIC] Hi, welcome to the CX Tipping Point podcast. Citizens deserve a government they can trust. This is where we bring together leaders from government and industry to reveal the secrets behind how we're transforming customer and citizen experiences. On the show, we chat with unsung heroes who work behind the scenes to improve the lives of citizens and government employees. We learn from the people who do it and the challenges they face. On your host, Martha Doris, let's do it. Welcome to the CX Tipping Point podcast. Today we have Jeremy Wilcox, Senior Director for Strategic Solutions for C3AI, and the former White House Digital Services Expert, an acquisition strategist, and a DHS, CX, and AI Senior Advisor. Welcome to the podcast, Jeremy. Thank you so much, Martha. Happy to be here. I'm really looking forward to this conversation and we've been chatting beforehand because we've known each other pretty long time, which I'm sure we'll get into. But you've got a range of experiences around service delivery and customer experience. And of course, that's what I've focused on for the last 10 or 15 years myself. And so I'm looking forward to bringing together that as well as your knowledge and expertise on artificial intelligence and how it can be used to improve productivity, improve the experience of the workforce and provide better services to the customers when they're accessing government services. So let's just start by getting your background journey to where you are today. Sure, Martha, yes. So it's been an interesting journey. I started out as an account manager in revenue generating roles. For the past 20 years, I've been focused on the federal space. And I was just thinking back to my 20s when I was meeting with chief information officers and the government, seal of executives to challenge thinking and read change. And that's where I was at forced to research going back 20 years. And as I think about the reason why we would have meetings, these 30 minute research meetings with executives. And then I think about the capabilities, not just of the internet, but with AI where we have these kind of pocket analysts that we can pull out to become smart very quickly. It just is amazing to see how I guess transformative and different technology has been over the last 20 years, but also there are a lot of things that are still the same. Yeah, that's my new best friend. All of our best friend, right? Yeah, I mean, so I spent about 10 years at forced to research. And then I found myself. And actually, you and I cross paths, I want to say around 20s, 20s, 10. It's probably spent over 15 years. And so before I guess before I get into AI or before we talk about things, there's a date that sticks out in my mind of like one of the two dates that when I think of Martha Doris, it was, and I looked this up because there's a news article about it, but it was March 22, 2012. And this is a time at GSA. I was the account director serving GSA. You and I had a meeting. I don't know if it was on the calendar or we planned to meet, but I was there all day. And it was at one constitution. So this was known to folks at GSA as the swing space when 1800F street was being renovated. But on this particular day, I had a series of meetings. And so I'm in the building late afternoon and then the alarms go off. And so first there's a lockdown. And they're telling everybody to stay put. So we go to the windows and we're looking outside. They're rolling out caution tape around the block. And they're shutting down the streets. And then they come through and they say, "All right, everybody needs to evacuate the building." And so we're not able to take our cars, you know, mines in the garage. So there are a few hundred of us at least standing outside on the sidewalk, a block away from the building. And then, you know, the rumors start swirling. There's been a bank robbery. They're, you know, not only was there a bank robbery and the, you know, the perpetrator was armed and dangerous, but someone had called in a bomb threat. And so it was a combination of the two. And so we were all, the whole area was on lockdown. And we're standing around for about an hour. And finally, my boss and I just said, you know, let's call it. So I walked back and I couldn't get to my vehicle, but I was able to talk my way to get to my trunk and pop the trunk and get my duffel bag, which had my gym clothes. I put my backpack and my laptop and everything in it because I wasn't able to take the vehicle. And so there I am. I'm just going to take a cab home. This is a pre-Uber days, right? So I'm hustling in a suit with a heavy duffel bag. Pustling up, you know, towards some union station. I'm a big guy, so I'm sweating. And I run up to a cab and open the door and throw the duffel bag in and I hop in. Oh, geez, I know I took, now I get the duffel bag reference. Yeah, so I throw the duffel bag in and I hop in and I say, go to, I said, take me to Dallas because I lived five minutes from Dallas. And he looks at me, he looks at the bag. And on the radio, they're talking about this bank robbery, you know, suspect armed and dangerous. And he just looks at me and he said, I said, and I just said, I'm not the guy. Trust me. And he goes, you sure about that? I said, no, no, no, I'm not. Just take me to Dallas. So he starts driving me and, you know, it was weird because he starts opening up and telling me stories about DC from, you know, his rough upbringing on the streets of DC. And he's like, you know, telling me, you know, like he kept saying, like, you know, if you did do something, we're cool, we're cool. You know, the more I pressed and said, I'm not the guy, you know, I guess that's what a bank robber would say, right? We're heading, you know, and this is 30 minutes in. So we're going through Tyson's corner and then my blackberry buzz is in it's you and you're basically saying, like, are you still here? You know, we're back in the office. Can we meet up? And so at that point, I make a flash decision. I'm like, you know what? I'm going to turn this thing around. So I basically said, you know, hey, I'm in a cab, but I'm going to turn it around. Be there, you know, I think like, you know, 530 or something. And so I tell the cab driver to turn it around. He takes me back. But, you know, again, he's telling these stories that get crazier. And I literally took a picture of his, you know, his, I don't know, his, his license or whatever. And I sent it to my wife and I just said, hey, if anything happens, I'm in this cab. But we get back there, you know, still not the bank robber, but I get back to GSA, you know, to the swing space and we meet up. And, you know, what's interesting about this story is, is that we, the reason why I wanted to meet with you is because we were going to talk about the Forster proposal. There was a proposal for the customer experience leadership board. And you ended up becoming the first federal federal employee on that board. And so I always, I like to tell the story that, you know, it's like the butterfly. I actually never know, you know, what it might lead to. But, you know, two years later, the word, you know, you call me up and say, hey, I've got an idea, can you get George Colony to come down, George Forster Colony, Forster, to meet with Dan Tanger-Lini? And I said, sure, maybe, why? And you just said, trust me, I have an idea. So I went with that, you know, that idea was yours, Martha. And so I knew how to get George into the room and I did. So George comes down and see me, you, Dan, who was the GSA administrator, George Sonny, Hash Me as the CIO, and, you know, a few senior advisors. And so I'm sweating because it's one of the biggest accounts of Forster at the time. And, you know, we're going toe-to-toe with other big firms and we were holding our own against the other firms. And so George, we roll in and he asks for a white board and three sharpies. And I don't know if you remember that, but they had to find one. And so, you know, we're a tech research company, but George is a master storyteller and wants a marker in a wall, which I thought, it's like, honestly, it's like sometimes the most effective thing is like a log, right? And so, you know, as we're talking, we're meeting with Dan. I had prepared all these different questions and at one point, there was one part where Dan turns to me and I guess this was something that Dan often said, but he just said, you know, Jeremy, how long have you been working with us in seven years? Sorry, and he said, well, then if you were me sitting in this chair, what would you do? And so the one question I wouldn't, you know, it didn't anticipate. And I guess whatever I said helped push things along because four to six weeks later, I get this call on your static and you're like, we did it. And I didn't exactly know what you meant by we did it, but the first, you know, CXO position was created at GSA as a result. And that snowballed really if you fast forward to, you know, the veteran's experience office at at the VA and other agencies. Well, the federal student aid already had it had one and so did XM Bank, then us, the VA, and it started expanding from there. But I had been interested and involved in learning about customer experience for a couple of years before that happened. And yes, and I wanted to create a customer office in GSA because my opinion was, you know, with the, being an agency that as the current administrator talks about it as the engine for government, and all the agencies are our customers and all the different segments of those agencies. I just felt like once I had learned about it, I just knew that it was the way to, it was a business discipline that would help not just GSA, but all agencies. And so, yeah, we were off to the races there. Yeah, well, I mean, and I was just, you know, glad to be a part of that. I guess, origin story with you, Martha. It was, it's great reminiscing and looking back, but also, I know, you know, you have an event every fall. And being in the room with hundreds of people that are, you know, a part of that, you know, is impressive to see, you know, the community come together and, you know, celebrate the wins. A lot of unsung heroes. So, thank you for your-- Thank you. Yeah, I'll, the service to the citizen awards are, we just announced them this week and they're going to be the, the event this year, September 18th. So, this is our ninth year. Yeah. I've been retired to, over 10 years. Right. Yeah. I remember that. And so, you know, after, I guess, going back to, you know, how I landed where I'm at, but I mean, you know, after spending 20 years and sales, I ended up entering government. I applied-- I'll say this like, you know, I kind of applied on a win one night to the US Digital Service. And part of it was maybe a little bit about, out of frustration of, you know, just certain things that I hadn't, you know, seen done. So, I said, I'm just going to join government, see if I can, I don't know, become part of the problem or, or, or tackle it from the other side. And I submitted a one-page resume. And because a two-page resume would have cost me money if I, you know, subscribe. One page was like, you know, free and it was one of those services that, if you did more than one page. So, I just said, okay, I'm going to put one page through and next thing I know, month later, I get a phone call and they want to interview me and go through every round and get selected. And, yeah, so I spent 49 months with the US Digital Service. So, served in the, you know, Trump administration as well as Biden administration and then stayed on in government, it was convinced to continue serving as a public servant for another two years and served with the Department of Homeland Security in the, the Office of the CIO in the customer experience record. So, and that's where I really got into artificial intelligence and AI. Well, so if you're describing, you know, I want to do some like foundation setting around artificial intelligence and like, generative AI, agentech AI and the experts in AI throw a lot of terminology around. So, can you level set like, what are the major types of AI and what's the, what do they stand for? What do they, how can you use them? Sure, yeah. So, I mean, I'll try to keep the label simple. You know, in terms of AI, so on November 30th, 2022, I like a number, I guess, thousands or hundreds of thousands of people, I can't remember how many people, but Chatchy PT 3.5 dropped and it's the first large language model that came out publicly, L-O-M, and it's a generative AI model. And really what it does is it helps create, or summarize, or explain, or draft, or translate, or search knowledge, right? So, you have these large language models that are trained on, you know, whether it's like the library of Congress or everything they can find online. You know, I think almost everything has been, these large language models are so large now that I think they've been trained on all the public information they can find. And it's really, you know, that's Gen AI. When you hear generative AI or Gen AI, or if you hear frontier models, that's the, you know, on the frontier, these are the leading edge models that you hear about. You know, a Gen-Tik AI is, you know, an extension of that, meaning you can take a goal or use tools, or, you know, have it reasoned through steps, and help move a process forward with human defined rules and controls. But, you know, the lines have been blurred to the point where if you use any L-Olem at these days, you know, agents spin up in the background before you even notice. So if you've ever done a search recently where you've asked it to develop a document or review something or what not, you'll see these agents spin up in the background and produce or connect to tools, workflows, and so forth. I know I asked at the other day for some information about digital services and customer experience people in off of these dates. Oh wow. Yeah, they gave me a spreadsheet, but they didn't give me all the information that goes in the spreadsheet. So I will be back. Yeah, and with these agents, I mean, they, we're getting to a point where, you know, you can direct these agents to almost work continuously, and they'll come back, and you can check in with them. You know, is that a bootcamp where they, you know, the model spun up six sub-agents and, you know, it kind of, I had my own, but all agile development team that was, yeah, building an application. I built a couple of applications and just for fun. And I didn't really get the end product I was looking for, but I had the agents running all night on my I think one of my sons gaming laptops. So, I mean, agencies are really, I mean, this administration's really pushing, we're really leaning into technology first, basically, and the use of artificial intelligence. And then there's, you know, agencies are also struggling a little bit in developing the skills and the people that have the skills to use AI effectively, and then also, how do you use it responsibly? I mean, what do you think how agencies should be approaching the use of AI or their AI strategy? Yeah, I mean, I, I think that, you know, there's three questions that I, that I think we should ask. It's like, what would, what work are we trying to improve? You know, what decision are we trying to support? Or, you know, what, what risk do we need to manage? I'll push back a little bit on AI strategies, just for the sake of strategies. And I learned this at Forrest, where it was like ingrained in us, you know, people process and technology. Right. You know, when you pick the technology first, so for example, if you pick the model first, or, you know, you pick, we're going to be AI first, period, without understanding the people, or understanding the process in the workflows, you know, things can, can kind of go off the rails. And so, you know, rather agencies really should, you know, develop a service delivery strategy, or, you know, a workforce strategy, or have a data strategy, they have all these things. An acquisition strategy and a governance strategy. And then you could say like, how should AI support those things? Right. Because if you don't, you kind of like, you end up with AI, you just end up with pilots. And so I think if you really start with say service delivery, you can find where the bottlenecks are. So the goal is really using the technology to, you know, help government solve a problem, speed up the effectiveness of the process, basically, as opposed to, I mean, I think there are some applications or some places where they're definitely looking at it as a way to be more efficient and not use so much, so many resources in a space. especially in like context center space where you could use AI to answer questions through chatbots or through AI that would eliminate the need for somebody answering the simple questions. But it's interesting, I know over at CBP they recently created a chatbot compass and they found in through usability testing that people preferred to chat with the chatbot rather than like interacting with it. And so it was it's really a great case of how the CX side of things and the usability testing actually ended up saving them a lot of money because they didn't roll it out in one direction and then find out everybody wanted to use it in a different way. So that's always and chatbots are pretty well have been used for quite a while and not across the government. I'm sure they're improving over time. But what about like, I mean, HR procurement, finance, all those kind of functions, you know, it seems to me that there should be ways that you can leverage some, you know, artificial intelligence to speed things up a little. Yeah, absolutely. I mean, you know, I think we're seeing AI, I think AI will eventually touch everything to some degree, right? You know, it's you mentioned chatbots and you know, I think back to my XM radio, you know, renewal, you know, we get XM radio or whatever, serious XM. Right. And then, you know, it's usually free on a new car and then suddenly you get a bill in the mail and you know, you're like, why am I paying for this when I have iTunes and right, you know, multiple subscriptions to different music and radio. And so, but but reason I mentioned that is the last couple of years I've they rolled out a, you know, AI agent and I negotiated, I shouldn't say this out loud, but I negotiated, I don't know, it was probably 70 or 80% discount, you know, with the agent and I and it was it was a lot easier, I think talking to an agent in that case because I just knew it wasn't a person. It just communicate my bottom line. Here's what I'm going to pay or, you know, I just had a similar situation happen. I called because I got I was going through my credit card bill and I'm like, oh my god, look how much this is getting. Now, I'm outside. Oh, go ahead. I had it on two cars and anyway, I called up, I don't even know if I talked to a real person or an agent, to be honest and I ended up dropping one car and, you know, yeah. Well, on the flip side, I when I thought I'd lost my wallet one time, I was very frustrated at, you know, this agent that just couldn't put me through to a live person, you know, to and it was super frustrating because it was a unique scenario where I didn't know if I'd lost it. So I didn't know if I wanted to cancel it or get a new one. Right. And I just want to talk to a live human. I've seen the same thing happen with, you know, pharmacies where somehow I got routed to a different pharmacy, you know, than the one I had called through the, you know, the agents. So for me, that kind of like speaks to, you know, there's a right and a wrong way to deploy AI. And I think starting with people and understanding the process and where things break down or where the friction points are in any process, those are, those are, you know, opportunities, right? Well, and you know, the reason I mentioned the CBP example was just to show kind of the integration of some of the customer experience, like strategies or practices that are used and the use of technology so that you're not, you know, again, knowing who your customers are and what that process is and what you're trying to solve kind of brings the two together. Given the other examples in terms of service delivery, I know we've been, we've been talking about that through some of our kind of communities that get together and talk about customer experience and service delivery and how agencies can can use AI. Yeah, I mean, you mentioned, you know, contact centers and I think that's a great place to start when it comes to service delivery because it really, it, it, it reveals, you know, where, you know, service can break down. People call because, you know, the website or form or a letter or policy or process just doesn't answer the question that they have. And so AI could help the agents, but it also can show leadership the patterns around, you know, what people are confused about, you know, where are they getting stuck, what policies create repeat calls. I mean, that's a big one, right? Re-set my password or, you know, there's a lot of different things where you get repeat calls, which processes create unnecessary burden, right? I mean, and, you know, when you think about a multi-channel strategy, do, does the contact center have access to, you know, is the left hand talking to the right hand? I remember filling out, I was trying to set up a grill at home and it was missing some parts and so I put in, you know, a, I filled out a form and I got a phone call and they couldn't, they couldn't see the form. So they didn't know the details. And so that was just, it was like a compounding frustration. There you go. That's another customer experience, best practices, connecting connecting the channels on the back end. But don't you think AI is especially valuable when you have a lot of data to, that needs, you know, analyzing to be able to take action on something. So if you, you mentioned multi-channel. So if you've got phone calls coming in and I know AI is really good in the contact center space in terms of analyzing call recordings, identifying root causes so that you can improve the websites based on what you're, you know, what people are calling about. But if you had that, if you bring the data together from the web, all the channels, right? Like the, the reasons that people are going into a, if they have in-person options and the website, the contact center, they're using AI can help them analyze it and prioritize things that they could fix. Yeah, most definitely. I mean, you know, one of the things that I would encourage is say, you know, don't start with fully automating decisions. So you could take all this data and look at trends and analyze it. But I mean, AI is great for, you know, summarizing that information, but also drafting responses, for example, right? Or, you know, looking at, or searching policy documents, pairing case files. So you could get 80 or 90% of the way there, like a jump start. But then having that human in the loop or human in the lead, right? To empower the workforce. And you mentioned, you know, AI is going to be integrated into everything. And I think that's an important point because if we don't, you may not realize it, right? But any time you now Google anything, right, it gives you the AI response first. And I don't even Google things the same way that I used to, right? Like I search for things. I asked the question that I would ask if I had gone on to ChatGPT. So then you get, I mean, I even went on to ChatGPT and asked it, I said, okay, I'm this age, tired from the government, been saving in my 401k, you know, what I'm going to take social security at this age, did, did, did, did, what, give me a strategy or tirement strategy. And it was crazy how good it was. Yeah, no, I mean, and if you, you know, the other thing is if you have, if you use Google, like Gemini, for example, and, and I think the default or it asks you to turn on hyper personalization, then suddenly it has access to a lot of your information, including your emails. And, yeah, I mean, to the point where there was a, there was a, I'll just tell this funny story. So, you know, a couple of my children, they work at Chick-fil-A. And my pleasure. So exactly my pleasure. Yeah. So there was, I think Gemini rolled out this new professional headshot feature, you know, with AI. And so I uploaded a photo of myself, just my, you know, headshot. And I was in my car, but, but I was not like in the parking lot of Chick-fil-A. But Gemini took the liberty of adding me, sitting in, like, wearing my suit, sitting in my car, eating a Chick-fil-A sandwich, and having a, like a milkshake, like a mango shake. I think that's the, that's the seasonal, which just dropped. And so I, I asked it why. And it said why? Because you, you know, would jump the gun. And, you know, I know you love Chick-fil-A. And you love those, you know, mango milkshakes. And I just couldn't believe it or peach. I said mango, it's peach. So the peach milkshake is out there right now. But it hadn't even, it was like this was on a Friday and I asked my daughter, you know, when is this seasonal peach shake coming back and she said Monday. And so, you know, in the back of my mind, I'm thinking, you know, is this, is this like AI product placement now? So it's like, at least they didn't put you in a Chick-fil-A outfit. Well, that's true. Yeah, sure. Yeah, it literally said, you know, it was saying, you know, you know, it took the inspiration of my, you know, what I thought, you know, my desires or, you know, my passions were and, you know, put me sitting in the heart of the lot eating. Yeah. Yeah. So, but I wasn't at Chick-fil-A when I put that prompt in either. It just put two and two together. Yeah, you know, and I think that's also where, like for me, that's, it's like when AI gets it wrong. So, you know, are you familiar with, have you ever, have you ever asked AI a question and then, you know, it comes back with a false sense of confidence where it doubles down, right? They call it a hallucination. I have asked a question about specific laws and it came back and it did not have the right information about the right law. Right. So, this is like one of those things that I would say is like a danger that we need to be aware of. And that's that, Lawsniac, this, I don't know if you saw the trend where, you know, there were famous people giving comments, been addresses and they would talk about AI and they get booze, you know, from the their circuit. Yeah, yeah. And then, Lawsniac says, you know, he talks about AI, he says actual intelligence and he starts talking about the people, you know, he said, you know, humans will never be replaced. And then ultimately, you know, I tend to agree, like humans are, we have to treat AI, like any other revolution, you know, the industrial revolution. And so whenever I hear people say, like we're going to eliminate humans, it's kind of like what's the purpose of life, you know, not to start going down that path. But really, it's like harnessing AI, you know, the power of AI as a tool. And then, constantly fact checking. So, for example, you know, if there's a domain, it can it can bring up to speed really quickly in a domain, you might not be an expert on, but, you know, it can still get it wrong. I think, I don't know which version of Opus it was, but I believe one of the latest models had a 92 percent, you know, accuracy, right? Which meant 8 percent of the time it was still wrong. And if you're not sure, then if you go with it, you could end up, you know, taking false advice. So, I just say it's like it's a great tool that can, in some cases, 100 acts productivity. You know, if it's if it's labor that's menial or repetitive, right? What about if you were meeting with the CIO, I mean, the CIOs are primarily the leads for service delivery now, except for in the VA. And I think Treasury might have their CFO at this point, might be their lead. But under the government service delivery improvement act, past and January of last year, I guess. So, they've designated leads in the agencies. If you were meeting with an agent, you know, with a lead, where would you tell them to start? I mean, a lot of agencies have chief AI officers too, chief data officers. Yeah. I think they're starting with their people. And they're starting with the information they already have, their strategies, their and the processes they have. So, for example, you know, not too long ago last year, I was a, you know, public servant, a federal employee. And so, you know, government employees are often burdened under complexity with too many systems or, you know, too many policies, too many documents, or too many approvals needed too many handoffs, or just not enough time, right? And those are all impacts on the workforce that I think can be addressed with tools, right, with AI. So, it's not about, like I guess, from my perspective, it's not about replacing people, but it's about giving developing or creating tools that can be a force multiplier for public servants. Well, that's a great example because I've heard of agencies looking at, okay, how do we use AI to give employees access to exactly what they need to do their job, right? So, they don't have to learn 100 things. They only have to learn the 15 that they need to learn because of the being able to automatically do that. And so, I know, especially like even in the context center space, AI can really help agents to be able to draft their answers, right, if they have lots of different places that they get information from. And I'm sure they need to give a sanity check on that to make sure that, you know, it's not, it is given the right answers, but that really should speed up the ability to answer questions when you get information from multiple places. Yeah, absolutely. I mean, you know, I don't know if it's like a data cleansing issue, but it's like, when, if you, if the data that you have available to you is like dirty data that, you know, is not reliable and you just put AI on top of that, you just have, you know, a lot more or faster dirty data, right? So, I think there's a kind of a cultural and change management shift that the government is undertaking or that the government should undertake around trust and that's, you know, employees needing to know if you roll out a new tool, what it's for, and what it's not for, right? Also, like what data it can use and then what decisions remain human decisions, you can assign it well, I'll give you an example and then also how, you know, how errors are handled and whether, you know, this is considered, you know, enablement or surveillance. So, I would say, you know, it's having a conversation with someone recently and they told me that we were talking about AI and they said they have, I think, 15 agents that they've developed or that they use, but this person works in IT and they said, you know, in the future it would be great if I could get an agent that would talk to, you know, the contracting officer and have their agent engage with my agent so we could, you know, resolve this transfer or, you know, to get a letter saying that I'm the core on my contract. So, you know, it's been a couple months and so in the future, I think we're going to see, you know, that friction be reduced and we're going to be able to develop agents that will be able to deploy and, you know, reduce time or complexity into bite-sized chunks. So, for example, if you need something signed, you'll send it out and come back. There's questions that'll go back and forth and in some cases you could give these agents guardrails where they'll be able to negotiate within certain, you know, within a certain space and be able to get things done to save you time, right? I think your point also about data, you know, when you think about, okay, if you're in an agency where do you start? Well, cleansing the data is one of the first things as well as looking at your business processes so that when you, you're not throwing technology on top of bad data and bad process, right? So, there's still a lot of pre-work to be done to make the tools as useful as they could be. Right. And if you look at a lot of the contracts that we're seeing, you know, awarded are, you know, specifically about cleaning up data, right? Making the data available or making it so that it can be used across multiple systems. Right. What about the impact on the workforce and change management? I mean, a lot of people believe that it's just a strategy for reducing the workforce as opposed to a strategy for making the workforce more productive so that they can, you know, focus time on, I mean, I know a lot of things I use AI for that I would spend days writing something that now I can spend, you know, 20 minutes for the most part. It frees me up to do other things. Yeah. No, I mean, you're right. I mean, there's a lot of things that we can use AI, you know, in the workforce to save time, reduce the number of handoffs, you know, drive better quality, improve response times like I mentioned with, you know, between, say, a core and a CO, just to deliver better overall. Employee experiences, but also customer experiences. And it's the, I had a conversation recently with somebody who said, that I think it was called at the laundry, like who wants to do laundry? If we could get AI to do digital laundry or whatever, any kind of laundry. - For me, it's absolutely. - Why would a human ever want to do it? Right, exactly. And I think by this point, if you haven't seen the video of the human versus the machine, when they're sorting packages, have you watched that? - No. - So I think the human was in the lead, sorting packages within the first day. But the machines can recharge and don't need to eat or sleep and really take breaks beyond charging their batteries. And so they can come back and kind of win that in the long run. But I think that we may find like a scenario where in the future where we're, we still enjoy some of the, I don't wanna call them mundane, but like some of the simpler tasks, 'cause I think that we're gonna face scenarios where only the toughest tasks are gonna come to us and all the things that were easier are taking care of already. And so that's gonna be an interesting future, right? Where can we achieve 100x productivity out of a human? - Yeah, because if you're doing like the, going 150 miles an hour all the time, you're not gonna have a break. So you're not a robot, right? Like it's interesting because, but that's a great point. What about you? - Yeah, I'll say, go ahead. - Yeah, I was just gonna say change management. I mean, how do you, how do you get your, engage your, your teams, your staff to teach them or give them access to ideas on how they can use the technology to help them in their work? - Yeah, I mean, what's interesting is, you know, the models are so great now, you can literally build your own trading program just by asking it. - Mm-hmm. - To teach you things. And so, the way I look at models, it's like a continuous conversation, right? And, you know, the model is growing up over time and, but can also be forgetful. The context window or, you know, you're having a conversation or you're chatting with multiple LLMs and, you know, constantly having to repeat yourself or maybe the model forgets. I know there are ways that you can optimize that, but I would say from a change management perspective, just starting with, anybody could start with, I don't know, you could, you know, ask a question. For example, you know, my parents were looking to move and I, you know, I asked, I think I dropped like, kind of their scenario into Claude and they're looking to move from Calgary to Ontario and it kind of came up with a checklist, right? So here's what you need to do in order to sell your condo or your townhouse and, you know, here's, here are the laws between provinces in Canada and here's kind of some steps you should take or things to watch out for. I guess you should just play with it in your own personal life and see what capabilities are, what the limitations are and then you can start thinking about how you can apply it. - Yeah, I mean, for me personally, I would say 99% of my AI activity on a personal time is just that it's experimentation to see, often, you know, not even edit something. I'll just see how accurate it is and kind of log into my mind and then think how I can use it, you know, in my work life. But, you know, you'll hear people say like, fail fast or, you know, just start using it and that's true, like you can start using it very quickly. And also, you know, share amongst your team members, teammates, people in your organization, but also-- - I need to use different, different, you know, providers. I tend to go to one or two, right? But to, I think they're different ones are better at different things, you know, whether, I mean, I've seen people, and I have actually done it myself is, you know, give it some information and tell it to do a graphic design for you, right? And you pop something out. - Yeah, it's amazing. I mean, you can literally create, you know, websites just speaking anything into existence now. You know, building applications, I think what we're gonna see is it kind of reminds me a little bit of the '80s, where, you know, one or two people could make an application or a video game or, you know, an Excel macro for their office and, you know, solve a very specific challenge or problem and get it done versus, you know, say, pay, you know, an expensive consultant for many, many months to, you know, reprogram the mainframe, right? - Right. - And, you know, on the flip side, you know, I remember pre-Fitarah when, you know, Shadow IT, sorry, Shadow IT was a big thing and now we're seeing, I think Shadow AI, right? Becoming an issue, right? Where, H, you know, I remember, you know, you could build something and break something pretty quickly, you know, and it would affect the general ledger of your organization. And with AI, having the guard rails and being able to, you know, if everyone can create an application or an agent, how are those agents and applications going to engage with each other, work together and scale, right? So, I think those are the challenges that are going to come in the future that we're already seeing organizations face today. - Well, is there anything that you want to share around with your knowledge of, you know, the use of AI in government that we haven't talked about? - I think we summed it up. I mean, I would just say that, you know, ultimately AI is not the strategy, like AI just for the sake of AI, but in the context of service delivery, you know, delivering services better is the strategy. And AI only matters if it helps the workforce and the overall public experience can be different with the government. And so I see AI as an exponentially, greater opportunity for us to deliver better service delivery and have better experiences, both internally with our tools, but also the services we provide taxpayers can be impacted in a much more positive way. - And the workforce too. I think there's lots of, you know, positive and negative impacts on the workforce, but stuff's gonna have to shake out. So, well, thank you so much. I appreciate your time today and hope to see soon and congratulations on all your accomplishments. And, you know, all of us that are in the federal, state, local, government service delivery are pretty close tight community. So. - Yeah, thank you for having me on, Martha. And I look forward to seeing you in person at a future event. - Thank you, all righty. (upbeat music) Thanks for listening to the CX Tipping Point podcast. We'd love to hear from you. Follow us on Facebook, LinkedIn, and tweet us at Doris Consulting. As always, we're looking for successes to share with other governments and to celebrate the excellence through the service to the citizen awards.