Latest / You've Been Heard / 434- AI Hype Needs A Business Case w/Tim McCrosson

Transcript

434- AI Hype Needs A Business Case w/Tim McCrosson · You've Been Heard

• 0:00
  Speaker A: Ultimately, the thing that sold it for my executives was I had my CISO in an executive meeting literally demonstrate in a live fire exercise how easily she could take control of the network from an adversarial perspective. And then I told him, I said, look, we have $60 billion under management for all the laws and that money is at risk, so we have no choice but to go and address this issue. When I did that, they saw how much risk we were facing, they brought the money to the table, we started the project. Just a few weeks later,
• 0:55
  Speaker B: After talking with IT leaders, I've come to discover one thing. And it's the same thing, you know, it's the same thing everyone else knows, whether they want to admit it out loud or not. Nothing gets done in any company without it touching it. Nothing. But here's the thing. When everything's on fire, when nothing's working, when the ransomware hits, when the whole business grinds to a halt and it is completely indispensable, you're the most valuable asset in the company. But when everything's running like a well oiled Tesla, you're a line item, a cost center on a spreadsheet. You've done the work, you've built the career, you've earned your seat at the executive roundtable. But having a seat at the table is one thing. How much nicer would it be to actually be heard? This is, is you've been heard. So Tim, let's start off with this. I mean if you have any deep thoughts or feelings right there, like, go for it, but I definitely want to hear what your first computer was and you how, how you got started out in this whole thing. Ed, give me kind of like the shrewd story how you got started in it and how you ended up where you are today.
• 2:02
  Speaker A: Yeah, so I started in it during the dot com build up building websites and eventually got into a contract building websites for housing and urban development of all places. And this is just during the, the dot com buildup years and which I think quite frankly is very similar to the environment that we find ourselves today with respect to AI. Right. So back then in the very early years of websites, you had these, I like to say flaming candelabras that were on like the different websites at the time. Ultra Vista was the leading search engine for good reason because they did it right. But Google figured out a way to more effectively monetize their search algorithm and actually make money off of it, whereas altavista just had the really good product. They didn't figure out how to make money off of it.
• 3:01
  Speaker B: Sad.
• 3:02
  Speaker A: So, yeah, anytime you do something well, people sort of ask you to do the next thing. And so it just went from websites, web development, cold fusion development, and then to project management. And so I was running redevelopment of big grant systems at hud. And for me, it was a pivotal moment. I was blessed to have the right supervisor at the right time telling me exactly what I needed, which was that I was completely successful as a project manager, but I would be completely unsuccessful as a human being because while my projects were all successful, delivered on time, on budget, nobody wanted to work with me afterwards because I was just too hard a driver. And so that was a really good lesson for me to learn at the time. And so eventually I changed my course. I changed. I very specifically changed how I handle people. And I went from being this sort of command and control director of projects to being more of a facilitator, collaborator, and especially a coach. And when I did benefit to. I found that I created a much more sustainable environment.
• 4:27
  Speaker B: Yeah, it's a hard lesson. It's kind of like.
• 4:29
  Speaker A: But worthwhile.
• 4:31
  Speaker B: Yeah, it's like I have expectations that need to be met for excellence. And all of you should also rise to the occasion of my expectations with me demanding them. Would you kind of learn like the leadership piece that you want people to actually want to join your parade and follow you? Not just get in the parade, because you have to.
• 4:56
  Speaker A: The. The inspiration, I think behind that, well, at least part of the inspiration behind that thinking is a Harvard guy named Daniel Goldman who did all the. The work on emotional intelligence. And he had, I think, a really good book at the time, which was something about developing leadership traits and trying to figure out the leadership styles. The people who are the most successful are the people who have mastery over wide range of leadership styles. Because sometimes you need the coach and sometimes you need the authoritarian and sometimes you need a facilitator, and sometimes you need somebody that's just going to collaborate with people. And so the best leaders are people who match that leadership style to what the team needs in order to ultimately be successful.
• 5:46
  Speaker B: And I think someone said this way to me one time, like, if you've got someone whose bucket is really full, they've always been crazy successful on the team. They've got the most wins. I guess, like, they're probably more apt to get the more frank, let's push you forward leadership role at that point versus the guy whose bucket is empty and is brand new and is making mistakes on everything. You can't be as Hard on him as you would maybe the. The guy whose bucket is full.
• 6:17
  Speaker A: Yeah. Well, for me, this whole process came full circle, like, probably seven or eight years later, when I was put in charge of managing a project that was very, very high visibility and had input from many different departments and agencies. And so there was the presidential memorandum that said, Basically, in 90 days, we're going to deliver this big thing. And so I was doing exactly what I had done, which was create that collaborative environment. And I'll never forget this. At the time, he was the CFO for the Department of Commerce. His name, Scott Quayle. Great guy. And he pulls me aside and he. And he says to me, he's like, look, man, I know what you're trying to do. You're trying to get everybody to rally around and build momentum, but quite frankly, we only got 90 days here, and we need somebody that's gonna get in there and actually start cracking the whiff so that we can deliver on time. And so for that, for me, it was like, oh, I had spent the last seven or eight years trying to be Mr. Collaborative. And I had really gotten away from and not used that old tool that I used to use, which was, let's go and set some direction and hold people accountable and try to get things done that way. And so it was fun and interesting to go and pull some of those old skills that I hadn't used very frequently back into my repertoire.
• 7:48
  Speaker B: Why did we use the word why flaming candelabras of all? Is it because everything was on fire and there was just on different things?
• 7:56
  Speaker A: You don't remember, they used to have these animated GIFs, or GIFs, depending on where you come from. Right? Animated GIFs on web pages, and they were like these candelabras that were animated, flaming. And by 2001, 2002, everybody sort of looked at that as, oh, it's the tackiest thing in the universe. Who would ever do something like that? But we all knew that we had. What is it? Excite webpages and things like that that we had built, and we put those animated GIFs on them.
• 8:26
  Speaker B: So let's go a little bit deeper into, like, if you had a moment in your career where it was the absolute hero, where if your team hadn't been there, the whole business would have been on fire and come to a grinding halt. Can you think of a very explicit story. If your Internet goes down and you're calling carriers yourself? That's not strategic. It. We help teams centralize monitoring proactive tickets and escalation across Every location without capex. Less noise, more uptime, make it boring again. Just go to you've been heard.com and answer the seven questions to simplify, streamline and save.
• 9:09
  Speaker A: Well, there's certain times, maybe not a grinding halt because it is a big government, but people sometimes get they complain about well why does the government do things like that? Why is it so inefficient? Why do you so much fragmentation. Back in my time when I worked at the White House for one and we had question they had thousand data centers and thousands upon thousands of separate acquisitions for the exact same software. I remember one of the worst cases was mobile devices. You know at the time it was at and T Verizon T Mobile as it maybe Sprint. And so at one agency we had more than 4,000 separate contracts buying mobile devices just at one department. And it's like why is the government so inefficient? Why don't we go out and achieve those economies of scale? And I have an answer the reason why, and this is my sort of you've been heard moment or story because none of the CIOs during those times had the authority to actually put their foot down and say no, I'm not going to wish you on that. You're going to have to buy it in a manner that's consistent with our enterprise architecture. You're gonna have to buy it as if we're one government for once.
• 10:28
  Speaker B: How would that work though? I mean that's the whole like the answer to your question is the government. Right. But it's a two party system that switches every however many years and all of the personnel and everything inside and then everyone's. And then we've got this bid process and something called like a service disabled a tribally owned Alaskan status purchase program so that I can bypass a bid process so that I can. I mean it's like the whole thing itself is like that's why I really just don't do work with the, with the Fed. I will help people be get CMMC certified so they can sell to the Fed. But understanding that process is like it's. Is this what we're talking about though? Is that what you're saying? Is that why we have all this
• 11:19
  Speaker A: like it is a portion, right. So I was always career Fed. Like I don't look at it as being team red or team blue and it is black and so it doesn't matter what party you come from, you can find good things to do from an IT perspective regardless. But the problem that we needed to solve was in terms of budget authority.
• 11:41
  Speaker B: Uh huh.
• 11:41
  Speaker A: The CIOs that the departments did not have the authority to say this is what we're going to buy. That authority was dispersed among the programs and the bureaus. And so we put policies then in place. The most significant, the federal CIO at the time was Steve Van Roeckel. And what we did, it was called the CIO authorities memo. And I remember it distinctly as if it was yesterday is going to be very wonky. But it was OMB memo M 1129 which was a watershed moment in federal IT because that was the one where we said if it is commodity it, so infrastructure, types of things, laptops, network, mobile devices, software. If it's commodity, the CIO of the department has the authority to manage those things. And that was the first time in 2011 that anybody had ever conveyed that type of authority to a cio.
• 12:51
  Speaker B: Does that work now? Are people savvy enough nowadays to know though, to know about APIs and webhooks and interconnected software and stuff like that? Could that screw up different solutions that might be needed across the government? Well this is kind of the thing that. So this is if we want to really get deep into the psychology of how high level CTOs or CIOs make decisions. A large majority of CTOs and this is what I've found based on 22 plus years of seeing how people make decisions. Well first of all, it might be a political decision. It might be someone might make a decision because so and so's son works at at Microsoft versus Zoom or something like this or Comcast or numerous other Adobe. Like we can buy our Adobe from 15,000 different ways but we're going to buy it this way because so and so's son works over there. That's one way. That's a political thing. The second thing would be it's just safe. It might not be the best decision, but we're going to make this decision because I know like the old saying, it's like no one ever got fired for choosing IBM. So was it IBM at first and then it was no one got fired for choosing AT and T. And what does that mean? It means that people make decisions not based on what's the best possible decision technology wise for the company. They make decisions based on what's the safest. Now nowadays I would say it's almost the opposite. People make decisions based on fear of missing out. If we're going to talk about, like you said, AI and stuff like that, people are making all kinds of stupid decisions and wasting money. And I don't know how that's any different from some of these other ones. But I guess what it is is it's people don't always choose the best solution. They choose the solution that they can defend.
• 14:40
  Speaker A: I would say bedrock policies in at least in the federal space is a collection of regulations called the Federal Acquisition Regulations. The FAR and that is intended to govern all of those acquisitions so that we avoid nepotism. Right. So that we avoid favoritism. Right. And so it's. You can't award. And I would have to say that most people, there are very, very specific notable exceptions to this rule. But most people absolutely want to make the most efficient and the most economical acquisition that because honestly let's saw it like everybody else is. I'm not spending my money, I'm spending the citizen's money when I make an acquisition, when I go and put a contract out there.
• 15:37
  Speaker B: So it's good to know that the technology guys care about this stuff because I saw money get thrown away in all kinds of crazy ways.
• 15:42
  Speaker A: Well, yeah, I mean there's always stories about $4,000 toilet seats and $2,000 hammers and those types of things. Right. And those exist for sure.
• 15:51
  Speaker B: It might not be $2,000 hammers, but it might be. We don't know how many hammers we need for the job. So we're going to buy 200,000 in case we run out because we only have this one bid period during the year and I don't want to do this again. So. And I'm right in the RFQ or whatever. This, this is just what I've kind of seen. And then you've got, I don't know, there's a lot of stereotypes that get thrown around D.C. like the retired generals that everyone that are running around with their ponytail and they've got their driveways and you see their multi million dollar houses and stuff. I, I don't know, I stay away from the government space is such a crazy. There's so much red tape and stuff to cut through that it's kind of like this double edged sword where you have. It's like the policies are in place for everything that you just said that is good to make the best decision, the safest decision to fall within these reasons. But so then naturally people are going to how can I get around it? Yeah, like how can I partner up with again service disabled female owned tribally Alaskan owned company so that I can use this, a bid process to get through the Processes faster.
• 17:01
  Speaker A: Yeah.
• 17:01
  Speaker B: When I saw that, I was like, man, it just scrambles my brain. Can I just work in technology, please, and do what I love?
• 17:09
  Speaker A: I'm with you on that. And interestingly, I was talking to a couple of guys who used to sell hardware to the government and they made the comment, boy, oh boy, we always look forward to August and September because they ended. A government fiscal year is October 1st, right?
• 17:25
  Speaker B: Yeah.
• 17:26
  Speaker A: And so historically, agencies and departments would have money that would expire at the end of the fiscal year and spend it. What is the one thing? Use or lose. Right. That's the concept. Use or lose. Right. You can't carry over the money. And what is the one acquisition that you can guarantee, almost guaranteed is going to be a safe acquisition at the end of the year that you're going to be able to spend that money on? And it's almost always laptops. Right. We'll just go and buy laptops. And in my experience, I had an opportunity to work at agencies that were on opposite poles for this extreme. I worked at Housing and Urban Development, which was always woefully underfunded and never had enough money to replace the laptops that were aging out. And so it would not be odd for us to see laptops that were in service for seven years or more. So we would run our laptops until they fell apart and then you would have to beg to get a replacement. And a replacement laptop might not even be new.
• 18:30
  Speaker B: Yeah.
• 18:30
  Speaker A: Then I worked at a different agency that was very, very well funded and that agency replaced their laptops every 18 months because they could. Because if they didn't spend the money, they would get less money the next year. And so when I eventually went to the White House, this was one of the things that I absolutely was focused on cure. Right. Is because my goal from a policy, IT policy perspective was to put agencies on a diet to fix both HUD and this other agency so that HUD would be forced to replace laptops on a more normal pace consistent with everybody else. And this other agency would be prohibited from replacing laptops early. So I created a policy that said, look, you've got to pick your refresh cycle and then you've got to stick to it. And most agencies picked a four year refresh cycle, which meant the policy was that I get to replace 25% of my laptops every year. And from my perspective, it actually helped to sort of even out that distribution of acquisition from one year to the next. We weren't buying all laptops for the entire department in One August or so.
• 19:44
  Speaker B: There's probably a model for that. That's actually an interesting. Probably an AI model for that too. Like, what is the most efficient equipment, like whatever turnover process for staging. Because obviously you have to reimage things or however you're staging, sending all these things out, however you're. Your desktop design, all that type of stuff.
• 20:03
  Speaker A: Well, yeah, that's an interesting point, because five years later, I actually went to USDA as the assistant CIO in charge of customer experience. And so I now had 117,000 people who got all of their IT capabilities from me, including laptops. And at the time when I came there, we were imaging and deploying laptops in a completely fragmented manner. Right. So each office, USDA, has about 4,400 different offices. And so we had people. I had a team of about 1300 people who worked for me. And so part of their duties would be every month, some people would have to image and deploy several laptops. But when I did the math, I figured out that what I needed was the capacity to image and deploy 35,000 laptops a year. That's about our 25% refresh rate, which is what we chose. So because I had that policy experience from the White House, I recognized the issue here and created a new capability called the Enterprise Default, where we would have regular shipments from, in that instance, Dell that would come in to replenish. And our image time was when I first got there, it was more than eight hours to actually image a device intune and all of the capabilities, we got that down to less than an hour, 10 minutes. We can ship out devices the exact same day, which turned out to be critically important. During the pandemic.
• 21:55
  Speaker B: Yeah. Well, man, that's the pandemic. I had so many crazy stories from IQ Driven.
• 22:01
  Speaker A: I got the craziest story. If you need to consolidate networks, the best time to flatten competing networks is when everybody's working from home. Because I was able to consolidate seven networks between March of 2020 and Memorial
• 22:24
  Speaker B: Day and do like a pandemic. We should just do a pandemic for it. And this is just like a test. And then we can consolidate networks.
• 22:36
  Speaker A: 500 switches came into the loading dock, and it was easy because nothing else was working right. No one was there. The only truck that showed up that day was my truck from Cisco that had 500 switches on it. And then we started, we put together an assembly line and just deployed those switches to telecon closets over the course of the next two months. I was amazed.
• 23:05
  Speaker B: Yeah, that's actually kind of like that's like a fun project. That's where like it becomes kind of like, that's like a fun project.
• 23:11
  Speaker A: Well, I did go a little crazy on that project, I will admit.
• 23:14
  Speaker B: There's so many areas where it's where you can go find roi. As you grow into kind of like the enterprise role and you become more of a leadership, it's like, how do you let go of some of that stuff and how do you let other people take on the business of streamlining and bringing more ROI to the table? It's like, how do you really start to kind of like let go and trust your team to do that? Because if you start to think about it almost, you'll just be like stressed and won't be able to sleep at night knowing all of the endless to do list of things that there are in it. How do you deal with the stress? Because for any IT director, there could be easily 25 projects that people are asking you to do or 25 different areas like how do you really kind of grasp, get a hold of all that and eliminate things and really kind of wrangle the business together, Be on the same plate with the C suite, so to speak, on what the main challenges are and what challenges we need to kind of just say, hey, you know what, forget about that because it's not going to get done and it's not important right now or it's not important enough.
• 24:21
  Speaker A: Well, you have to invest in both managing up and managing down, right? And so when I look at managing down, it is, to what degree am I equipping and cultivating people who can take my job? That's my goal, right, is to put myself out of a job. Because I have built up people who can exceed my capacity, my ability. I was really, really happy 2025, I was the mentor, four different people. And all four of those people were promoted that year. Now, one of them was promoted because he was in the reserves, right? And he was promoted to captain in the reserves. But I still count that as a win. So four out of four. And so anybody who thinks that it's not worth investing in those types of scenarios, mentoring people and really investing yourself in that, you're selling yourself short. At the same time, you have to cultivate those relationships with the executives. And so at the same time that I was mentoring these people, I had a dilemma and I really needed the executives. At the time, I was rural development. Rural development in USDA is a big bureau. They make a lot of loans and I had A situation in which we had literally 40,000 vulnerabilities and we were never going to get, you know, there was, I would say before I got there, years of deferred maintenance. And it created a degree of technical debt that we were never going to overcome. And so with my support from my ciso, I developed a solution to get us around that, and that was to deploy a SASE. And so if I had a SASE, then I could obscure the 40,000 vulnerabilities that I was carrying on my network. And I needed $3 million from them in order to acquire and deploy that SASE.
• 26:35
  Speaker B: Can I take a guess on who it was?
• 26:38
  Speaker A: Sure.
• 26:39
  Speaker B: Cato Networks.
• 26:42
  Speaker A: Cato's great, but Cato, in fact, they're number one on Magic Quadrant. You were very close, but Cato is not Fedramp authorized.
• 26:54
  Speaker B: I bet you they are now after losing.
• 26:56
  Speaker A: Maybe I don't know that they are. Zscaler was the only Venom that had a Fedramp authorization. And I worked with them for more than a year to get that project going. Ultimately, the thing that sold it for my executives was I had my CISO in an executive meeting, literally demonstrate in a live fire exercise how easily she could take control of the network from an adversarial perspective. And then I told him, I said, look, we have $60 billion under management for all the laws and that money is at risk, so we have no choice but to go and address this issue. When I did that, they saw how much risk we were facing. They brought the money to the table. We started the project just a few weeks later. So it's managed both up and managed down. You gotta have both sides. Great people above you and supporting you, great people with you and giving you those brilliant ideas.
• 28:10
  Speaker B: How big of a sale was that?
• 28:12
  Speaker A: That was enormous.
• 28:13
  Speaker B: In other words, like, how much like how big of an ask was it? And the reason why I want to ask that is because I want people to know what's possible, give people hope.
• 28:22
  Speaker A: So I was the co chair of the Information Technology Executive Review Board, okay. And even though I was the co chair and could set the agenda, that was not enough. What we also had was a separate executive board. It's called the Enterprise Risk Executive Board. And that was made up of completely different people. And usually the risks that they're looking at are human capital risks, risks that the financial systems would have problems and not pay people appropriately and those types of things. And so it was really weird for the CIO of the agency to basically come to the Risk board and say, I have a risk that I would like to the enterprise risk team to consider. But then they did consider it and they prioritized it as the brand new number one risk facing the organization. And so with both the IT Investment board saying this is a very high priority thing and then the enterprise Risk board saying this is our number one risk, then it became something that the administrator absolutely has. And so yes, it's cultivate that support, work with your peers to help them
• 29:50
  Speaker B: understand at the end of the day, like how does one get approval for something that's not in the budget? Were you borrowing from the risk guys budget? Were they like, well no, we're in on this and we've got money for this even though it's not in the IT budget. It's kind of like we're sharing that together and like, yeah, we're going to make this happen.
• 30:08
  Speaker A: Well, federal budgeting is very challenging sometimes because you really have to be two years ahead in your budget cycle in order to get there. So sometimes the political appointees have reserve dollars that they can point to at their disposal and in this instance it was some of those reserve dollars in support of one of the programs that they prioritized.
• 30:33
  Speaker B: We will find a way. Yes, people wants to find a way, they're going to find a way. Just for everyone out there listening, both Zscaler and Kato are both close partners with us. So if anyone would like the extra special platinum support through Zscaler and Kato, then you can contact the backend program@infou'vebeenherd.com and we will make a red carpet introduction right over to the best team over there that gives us extra support in VP level access and premier pricing.
• 31:08
  Speaker A: Cato is supremely impressive. Just a couple of months ago I actually met the CEO of CATO at an event that they hosted down in Fort Lauderdale and man, talk about really being ahead of the curve both in terms of implementation timeline and usability. I will tell you it was 100% part of my three year roadmap to implement the CATO SASE capability. We weren't going to get there in year one or two, but year three documentation from that was what my target was.
• 31:49
  Speaker B: From secure cloud access to mobile devices, accessing the network to their. They're not even known as an SD WAN provider, but they have that feature backhauling traffic when we were in meetings before. I've had IT directors say if you take Cato away from us, I'm, I will quit tomorrow.
• 32:04
  Speaker A: It's the ease of implementation is the key that's probably what sets it in part. And don't misunderstand, Zscaler was great and I thought the Zscaler solution was implemented very well too. It was quite frankly as turnkey as I thought it could be. And then Kato's is even more so.
• 32:23
  Speaker B: It looks like they started their fedramp process in actually March of this year. So, yeah, still not ramp, but I don't want to sell into the Fed anyway. So anyone out there that needs Cato, you know what I mean? It's like I don't want to be in the Fed anyways. I want to help the people that are in manufacturing, health care and all those other spaces. But for all of you Fed guys out there that are working within the government, they have started that process and it looks like in March. So according to the news here. So, all right, what's your prediction? Where do you see us 18 months from. It doesn't have to be an AI prediction or something like that, but how do you see, see the IT leadership space changing? I think we're going to kind of level out of this bubble and realize that people wasted a lot of money trying to implement AI and a lot of time and starting to see that it's really more about implementing various different point solutions across AI that really make people's jobs easier. There's no doubt that it's, that it's helping, but I do believe that there is a level where you can go too far and it can become a kind of maybe a project nightmare or sinking debt and waste of time.
• 33:28
  Speaker A: So I have a couple of thoughts on this. The first is I've been working with a company that created, I'm going to say, thousands of microservices and then there we are coordinated and wrangled them together to. We were talking about the federal acquisition regulations earlier. Right. It's a, a very complex process that's intended to ensure fair competition for federal acquisitions. Right. But it has a lot of stops on that process. And I've, I've started working with a company that has used AI to absolutely short circuit this very complex process. And so you can go to their product and basically say, here's my requirements. Can you write a statement of work? Here's some information about the magnitude of what we want to do. Can you develop a cost estimate? Can you identify the terms and conditions that we should put into a solicitation? And then you receive proposals from industry. Right. Based on your request for proposal. And it's going to have both technical and a cost proposal. And then you can pour that in and say, can you evaluate these technical proposals to tell me where their deviations from my work statement are? And can you evaluate the cost proposals to help me understand to the degree and since I've been in many acquisitions, a lot of times industry will say you're comparing carburetors and bananas and chipmunks, right? Yep. Because if it's impossible to conduct a good comparison among the offerors, then it's sort of a crapshoot. It's in industry's favor to do that to the degree that the government can get an apples to Apples comparison is always going to be better for the taxpayer and government. And so by throwing it through the AI, you can actually get closer to that apples to Apples comparison and tell them the industry that hey, can you consider revising your proposal to be this many people and what is your range of rates for this? Because that's not consistent with what we're thinking. And so in a negotiated acquisition, you can actually get much closer to an apples to apples comparison than you ever could before. So that's one. And I think that AI has a lot of promise at the macro level, taking these very complex processes and making them maybe not easier, but more manageable.
• 36:25
  Speaker B: And dual, it definitely takes down the old man hour effect of data entry. And I mean I just remember as early as three years ago when you have to and back to your 500 replacing during COVID switch scenario. Yeah, right. Like I can remember doing network refreshes on old MPLS VPN backbones where we were replacing T ones and we were forklifting networks. And when you do that on a hundred location manufacturing or logistics company or something like that across North America and Mexico and Canada and then you've got manufacturing in China and stuff like this. When you start to actually do that back in the day it was a very manual process. It was email China Telecom and get pricing here and talk with an engineer to find out this and then price this and price call Cato so you can get around the Chinese firewall and you're manually putting all this stuff into spreadsheets and having assistants do this and that type of thing. That's on a very simplistic level, right? That type of thing. And then looking at fiber location maps and what's the fastest routes and how do you know that this provider is not just quoting another of the same provider? So you've got path diversity, there's all kinds of things. Right. That type of thing can take hours now. And for the IT director back in the day, it took eight months for the guy, for his partners that were helping him, it would take a couple weeks with a team of 25 people, all different vendors working together and quoting things. Now that can literally be done in like hours.
• 38:12
  Speaker A: If that's going to be revolutionary now
• 38:15
  Speaker B: is if it's accurate. That's another thing you have to have like be able to go through and verify accuracy. But it's easier to go through and do that and find the missing pieces and stuff and verify stuff and find little hallucinations here and there than it is that a human would have done anyways. A human's gonna make mistakes and keystrokes and move decimal points and stupid stuff like that, which is reminding me of inorganic chemistry in college. And the test that I swear I got an A on, but somehow I had put a decimal place in the decimal point in the wrong place and then I got a D. And that's the thing, right? That's clearly where it's going to. Very, very helpful. I guess the other thing would be is how well do you think my assumption is it's dumber than I thought. But how well do you think people really know the general user, how aware is the general user of AI and how to use and how to use various different LLMs and tools?
• 39:09
  Speaker A: I think there's a rapid acceleration in the knowledge that the general user has and I think that that's going to be a fun, like it's no different than the dot com knowledge continuum, right. And so people didn't know how to use websites and E commerce and then all of a sudden they started, right? And so you had these early adopters who got on and then eventually they coached the normal regular people, regular Joes and then everybody's buying stuff from Amazon and so then it's just the laggards who are, I gotta go into the store and buy the thing, right? But that's now a, a much smaller portion of the population compared to the people who are buying online.
• 39:54
  Speaker B: Oh, you have an idea? We should have AI searching Amazon for like crazy deals and then reselling them in another somewhere else or something.
• 40:03
  Speaker A: But I have another take on this as well. And I think AI is going to change the way or change our conventional wisdom on a very specific topic. So the topic that I would put out there is we have these notion that commercial off the shelf shrink wrap stuff is always going to be the cheapest, best value, right? And remember I'm in government acquisitions, right? And so when we think about software especially sort of big Ticket ERP software. Right. SAP those types of things, big software packages. The, the conventional wisdom has always been off the shelf. Get it off the shelf, go and implement a vanilla solution for your general ledger, accounts payable, accounts receivable, all that stuff. Right? Do not customize. And the reason that you don't customize any of that is because it makes your upgrades over time cheaper, more efficient, easier to implement upgrades. I think that AI is on the cusp of flipping that conventional wisdom on its side because that make or buy decision that says go and get a COT solution now compared to the cost, but using an AI and vibe coding, a very customized solution to my environment that meets all of my requirements and doesn't give me the extra stuff that I never needed anyway. I think that, that when you look at the cost now, AI is driving the cost of that custom development solution lower. Is it low enough that it's in terms of total cost of ownership or lifecycle cost cheaper in the long run than the COT solution? I don't know, given some of the licensing costs, maybe.
• 42:09
  Speaker B: So thinking about how we've struggled with this ourselves and I all that it means is that I live with the same pain that everyone else lives with, at least from an IT directors and implementation standpoint. Especially when you say CRM, ERP and these various different. I can't imagine like a massive ERP forklift would be just so difficult.
• 42:28
  Speaker A: And so I was listening to your podcast a couple, a few weeks ago. You were talking about CRM and Salesforce, right?
• 42:36
  Speaker B: Yeah.
• 42:37
  Speaker A: So the question here was should I go and make that Salesforce investment versus go and create my CRM capability using AI capabilities and build it that way. And so this then is the $64,000 question of which is like a hybrid.
• 42:58
  Speaker B: Yeah, I think it's going to be someone that has the team, the money and the manpower to build the initial build with the hybrid CRM in it. So for example, we've been like wringing our hands over what CRM do we use for like ever. And it was do we just stay with Go High Level? Because we've had it forever and we've been getting it free agent who's like a partner with us that we do trades with like that type of thing. Right. We're like, well case feels kind of like that Go High Level, which is the other one was the white label. So Go High Level has like everything that you've ever want in like it's basically like they said, they have a team of developers, like okay, just build it all we're gonna just throw it all in one thing and it's Legos. It's just. But then what happens is the interface becomes kind of like an administrative nightmare for some people that don't really know, kind of go high level. If you don't like the gui, you don't like the gui. It's just like, I really want to love you because you have absolutely everything. Like literally everything. Yeah, Texting, webinars, like, I mean, integration with Zoom, every API. And then like you've just got everything that you could, you know, you want. Oh, yes, you can do that. Yes, you can do it. But the GUI just becomes so gooey. It's like, do you want the iPhone or do you want like the BlackBerry with the flip out thousand buttons? It's just like, I want the iPhone, but I would have to go do all these other things. So then it's like, okay, so then I guess we could build Salesforce and we could have this API and that, blah, blah, blah. Well, there's a CRM out there called, I think it's 20 CRM. I think it's called 20 everyone Google this. It's just free advertising for these guys. 20 CRM. I think they're making insane growth right now. Yeah, 2020 dot com. Pretty good domain there. I don't know how I got that one.20.com. Yeah, build your enterprise CRM at AI speed. So they give you like that out of the box, vanilla CRM, like right out of the box. And it's gonna like, pull in contacts and email. It's gonna give you a lot of stuff out of the box. And then they've just got a really nice AI interface that will build all of the bolt ons and it's really, really customizable. And I think you could probably pay a developer or somebody or someone else's good with AI and a developer to kind of like build out your enterprise CRM pretty darn quickly, customize the way that you want it, like the way that you want it, the way that your end users want to use it. Not the way that we have to kind of change Salesforce to make it work with. You know, I mean, we kind of have to have these dispositions this way and the dropdowns this way. Like, no. So we kind of went down that route and we were like, ah, we, we had to kind of invest our time and then we have to be very careful because we're a podcast startup and we have to just kind of be very careful with how we invest our Time. That could be like, another really deep question, right? Like, is this the biggest challenge right now? The CRM? Is that really the biggest challenge right now? Like, how I put Tim's name into the database and make sure that, you know, he knows that we're doing a roundtable this month. Like, is that the biggest challenge? Probably not. So that's kind of why we. Right, but it's like the shiny object. Hey, look what we could build. Look what we could do. And then we could do this, and we could do that. We could do this. And that's the kind of, like, there's this constant mental battle, like the CTO and the CEO at the same time, so I can create my own problems. So as I think one of the keys to being a really good IT leader, you tell me, is how well do we know business? And how well can we keep our own people that are asking us to do all kinds of crazy crap and projects all the time on track on what's actually the actual vision and mission of the business? Because IT guys, as much as they say we're the department of no. Or whatever, we're actually kind of the department of yes. Like, yeah, well, I could do that. Yeah, that'd be cool. Yeah, I could do that. I actually don't think we're the department of. No. I think we should be like the department of hell no. All the time. You know what I mean? We should be the department of. I'm trying to remember who said it was like, yes, but no. It was like, yes, but first, is this a major challenge of the business? How does this apply to the main challenges of the business right now? What's the ROI going to be on this? It's kind of like the AI roadmap, right? Like, do you know what AI is? First of all? Do you have a steering committee? Do you know what AI is? Who's on the steering committee? Probably should be the most important people in the company or some of the most department heads. Two, what do we actually want to solve with this? What are our major problems? Are these actual problems? Okay, can AI solve that? Maybe. Is the data in alignment and cleaned up? Step three for that, step four, guardrails and security. Is this gonna leak all kinds of crazy data or is this gonna screw us up and everything like that? And number five, is it worth doing? Is there an ROI to it, or should we just hire another person for data entry or something?
• 47:47
  Speaker A: I mean, what you're hitting on is the business case aspect for these types of projects, and everybody has sort of AI fever and there's a rush to go and not be seen as being behind in terms of AI development capabilities. But I think that the IT leaders responsibility is to paint a picture and show everybody what the North Star is. This is where we're going and we're only going to do the things that get us to this direction because we could go left and right and down, but we're going to up in conjunction
• 48:28
  Speaker B: with the C suite. Like we are a team a hundred percent.
• 48:32
  Speaker A: The IT strategy is born from the overall organization's strategy. You tell us what you want to achieve as an organization and we will develop an IT strategy or IT roadmap to deliver that.
• 48:51
  Speaker B: You know what's actually bothering me, which actually I'm starting to rate people on nowadays, is I've looked at some pretty big vendors as of recent. So I'm talking like for me as a small business investing a hundred thousand dollars and that's a lot of money for a small business, right? So I'm looking at only the best. What's really interesting to me is I'm evaluating the best based on are they going to tell me they've got this great AI product? And that actually makes me nervous because I've gone through the motions of vibe coding things, of using AI in the business. We have our entire AI driven process for producing the podcast and. But the most important thing is that the human aspect stays intact so that we're actually using the real words of humans. And that's the beauty of it's kind of this weird because it's a language model, right? But that's the beauty of how we don't use a language model by using the actual language. It's really this weird, deep psychological thing. So it's like if the language learning model is based on language learning, for anyone that really knows the deep understanding of the algorithms of how an LLM works and everything, it's pretty wild only to come full circle and say, I want to use what Tim says exactly. Not what the LLM comes up with or its ideas of how Tim feels. Because people can, they're starting to very quickly realize, oh, it's another AI written email. I'm so sick of this slop. As nice and flashy as it sounds, I'm sick of hearing unlock. I'm sick of hearing. Everything's like, it's the great unlock. Like, why are you saying unlock? Like everyone's saying unlock nowadays. And I'm like, dude, what's going on? So I'm looking at evaluating this really really high level vendors and I'm kind of evaluating them on like, are you not using AI? Because that might actually make me feel comfortable that there's still a human in charge. But everyone has to at the same time say we've got this great proprietary AI thing that does this and it's like this weird double insecurity thing. Do you understand what I'm saying? Do you get that feeling a little bit?
• 51:10
  Speaker A: For me, I think the key is technology is technology. Let's make sure that we keep technology in that lane. But it's the relationship with the people who run the technology. Right. Are you. Whatever company it is, whatever organization it is, big or small, do you believe that they are going to help you and are you building that long term relationship? Right. We were talking about Salesforce earlier and I'll tell you, I have a significant affinity for people as Salesforce because quite frankly, yes, there's a lot of AI behind the scenes and there's a big army of people, very qualified people. The people that I dealt with on a day to day basis.
• 52:02
  Speaker B: Yeah.
• 52:02
  Speaker A: Were 1 million percent invested in what we were achieving together as a team. And so for me that counts more than probably anything. Right. Is how invested are they in understanding my business and then bringing the right solutions that solve my problems today, I think is more important than anything. It's the relationship.
• 52:29
  Speaker B: Yeah, for sure.
• 52:31
  Speaker A: Yeah.
• 52:31
  Speaker B: So we can't let AI take over the relationship. That's the one thing we can't. We can let IT systematize or streamline and make things better. But my prediction would be, which is probably different than it was again six months ago, is that people are gonna get tired of certain aspects of AI really quickly. I wouldn't want to have it take over customer service entirely. No. Aiding customer service. Yes.
• 52:57
  Speaker A: I mean, we talked about business cases earlier, right. So I had what I called a significant AI business case when I was at Rural Development. So a lot of people don't know this Rural Development, they actually issue single family housing direct loans. There's like 170,000 people out there in America. Then when they write their mortgage payment every month, they're actually paying usda. And so we are servicing all of those loans and occasionally those people want need to call in and talk to somebody. And so my AI use case was I want, if Joe from Rural America is calling in, I want my AI to be able to look as that number is calling in to know before the agent picks up the phone and says hello, what is going on with that person's account. Have they missed a payment? Are they late? Do they have a payment due? Has there been a disaster event that has affected their geography? And so we are equipping that customer support agent with the right information before the phone call even starts. So the metric that I was focused on was decreasing handle time for that agent. I want that agent to be able to address the person's need as quickly as possible, get that person off the phone so that they are happy. Because nobody wants to call in. You don't want to do it, you're forced to do it because you have a transaction or a business need that can't be otherwise addressed. So therefore this is the last resort. You're calling in and you want to be off of that communication as quickly and efficiently as possible.
• 54:48
  Speaker B: Being a contact center expert and ex telecom guy, that's actually one of the best places that AI assists very, very well as in contact center. Even from an automated agent standpoint, that's where it's kind of like helpful, like appointment reminders and things like that. Even outbound calls. Like I'd be totally fine with understanding like a really well voiced AI agent calling me and letting me know, hey Phil, I just want to let you know your appointments tomorrow, are you still on for that? Could you just press 1 or 2 or something or just speak to me, just speak to me, you know what I mean? And is there any problems with that? And then send them to a live agent if they need to or like you said, calling in and being able to pop a screen feed an agent, being an AI agent assist to the live agent of Sally's here. Like you said, there was a hurricane or blah blah blah and she missed her last payment. And pull up all, pull up the account and all the screens and feed some stuff along with sentiment analysis and Phil, you're speaking too fast and too loud and it's like I always get the feel you're speaking too fast on my sentiment analysis.
• 55:53
  Speaker A: I mean the only thing I want, I'm okay with a human being out of the loop on is and I'm sure as a service desk parsing this Monday morning password resets. Just go ahead. We're going to get a 10 or 20x in flutters of calls on Monday morning. Yeah, yeah, just they're, almost all of them are password resets. Just let the AI reset people's passwords do it.
• 56:19
  Speaker B: There's like the meme that I shared the other day on like the calling someone and helping them set up their two factor authentication or something like that. I Mean, in reality, why not? And this is something we did on a roundtable, and I'm spending extra time because this is just fun, but we did a roundtable event on creating your company's cybersecurity villain and your company's cybersecurity villains. Really more like a mascot that people learn to love, but he's the cybersecurity villain's whole purpose is to, like, fish you and cause other problems inside the company and create all these issues. We could create a cyber AI mascot. You could have, like, an AI mascot that people actually bond with. That'd be interesting. It's an idea.
• 57:01
  Speaker A: I worked at CISA for a couple of years in cybersecurity, and what you're talking about are Personas, right? And so, yes, there are Personas. Is it state actor? Is it a crime syndicate or crime organization? Is it single person with a ax to grind? And so just Personas has different motivations and different capabilities at their disposal.
• 57:28
  Speaker B: All right, last question.
• 57:30
  Speaker A: Yep.
• 57:30
  Speaker B: If you could go back and tell yourself one thing you're going back in time and tell yourself one thing on day one of your IT career, knowing what you know now, what would it be?
• 57:40
  Speaker A: Oh, man. There's a big fork in the road that everybody in IT will likely encounter at some point in their career. And that fork is, are you going to make a choice to stay on the technical path, or are you going to turn and go down the management path? For me, I clearly chose the management path, and I loved it. Right. Cause I got to be ahead. There's a part of me that. That is remorseful. I mean, nobody's cold coding in cold fusion anymore, but that I'm not more on the vibe coding end of that spectrum. And actually being a developer, I kind of miss those development days and the creativity that comes from that. The excitement of. I mean, ultimately, it's. You look back on what you've accomplished in a day and you can say you can measure it. We didn't have this capability when the day started, and now we do. And that's amazing. Right. Whereas when you're on the management side, it. It's a completely different metric in terms of how you measure success and progress.
• 59:03
  Speaker B: It's definitely more emotional roller coaster. There's more of your soul out on the counter. There's probably more sleepless nights. And I think from that standpoint, to applaud you, you took the more difficult path, whether you knew it or not, because there's days where you're just like, I'm a bad human. I need to do better. There's just days where there's just, like, very difficult days. Whereas I know that there's plenty of times where I've sat on boards or I've been part of an administrative committee or something like that. And I've just said, hey, guys, here's the thing. I'm much better at taking orders I don't want to deal with. Especially when you're in a nonprofit or something like that, or you're volunteering in the community. Sometimes you see things you just don't want to see, you don't want to deal with. Like, it can be very, very difficult. You don't want to know everybody's personal life. You know what I mean? You just take more emotionally on your shoulders, and I think, again, to applaud you, that you have to agree to do that and sign up for that. And you have to really care and really want it from the emotional standpoint, from an emotional intelligence standpoint. Whereas it's easier to be the arrogant coder and sit in the room, be like, I could always do his job better. Let me go back and, like, talk about how, like, you know what I mean? That's how I see it. Did I hit it correctly? Like, someday I just wish I was back sitting in the bullpen, closing tickets and I don't know, doing whatever, fixing things and playing with the machines.
• 1:00:29
  Speaker A: There's a part of me that wishes that I could just be on a red team, right? I mean, red team, certified echo ethical hacker people. They have. I think they have the best job, right?
• 1:00:41
  Speaker B: It's the dream, kind of.
• 1:00:43
  Speaker A: I mean, you had almost no responsibility. Your job is to find errors in stuff that everybody else has created. And you get to be as creative as you can possibly be to gain unauthorized access into stock with the intent of creating a safer organization. And the fork that leads that direction is awfully interesting, too. And so I miss that.
• 1:01:13
  Speaker B: At the end of the day, everything is kind of a. The grass is always greener on the other side, right? Even if you were there, like, that company could be sold or something like that. But I guess the question then could be. Is, like you said, there's like, two. Like a fork on the road. Like, there's a third option. There's a third option, which I always ask people, like, what's the end game for it? The third option would be start your own company. The third option would be create your own thing, right? But then that comes with a whole nother host of prop. Believe me, trust me, it does. It Comes a whole nother group of sleepless nights. It's like the cook that's always loved cooking his entire life, or she's always loved cooking and baking or whatever it is her entire life. Right. So I'm going to start a restaurant. Bad idea. You're gonna quickly hate cooking. You're gonna quickly hate all of that stuff. Unless you're really driven and have, like, a real mission and purpose behind that. Because then you're gonna be managing employees and ordering paper products and timing, and then there's orders and there's fulfillment and there's all these other things that have nothing to do with your passion of cooking. So, yeah, if there's anyone out there working in the government space, whatever, it's the Hunt for Red October, whatever it is, that wants to hire Tim to just sit in a dark room and, like, break into places. You could just join the dark side, too. And we've seen a lot of security guys kind of, like, be underappreciated. When I saw in the news the other week that there was, like, this co conspiracy inside another company that led to a ransomware attack and the CISO was collecting on ransomware, it didn't surprise me.
• 1:02:40
  Speaker A: Wow.
• 1:02:41
  Speaker B: It's wrong. It's absolutely 100% wrong. And that person should go to jail.
• 1:02:46
  Speaker A: I knew some of those guys on the threat hunting team over at cisa, and you know what? They hats off to them. They do great work. Right. And so we need everybody. It's all hands. Everybody's got a role to play in making this whole thing work.
• 1:03:04
  Speaker B: Yeah. Sometimes I dream of going back and starting over again. Something. And then I'm like, I'm going to be 50 this month. I just need someone to remind me that I'm probably really close to death. And then I'll just spread the word. Everyone, you've been heard. Let's go. What it's about is the other side of it. Leadership and really making sure that it that has a seat at the executive roundtable is heard and that your story is penetrating the room in the company when you're not there. So what does everyone think when Tim's not at the table? What does everyone think when Tim's not there? Are they, like, worried that we need to bring Tim for this? Right. Versus that sales rep is a really great idea. I think we should buy that CRM and. Yeah, no, Tim and his team, they'll implement it. Yeah. Let's go. You know what I mean? Like, there's so many things that were missed. So that's the mission. So thank you so much for being on the podcast and coming in and sharing your years of knowledge. Any final words or thoughts?
• 1:04:03
  Speaker A: Phil, thanks so much, man. This was a great conversation. I and I. I really enjoy all the good work you and your team do. Like, I'm getting these stories out because I'd spend days listening and driving on in my car listening. And there's something to learn from everybody and a little bit of inspiration in every story as well.
• 1:04:23
  Speaker B: Yeah. And if anyone out there has any suggestions for things that you'd rather hear, maybe we should break up the show and have some different sections or something like that. Like, I don't know, AI Flops of the Day or something crazy that my CEO brought to my desk today, which we probably cannot air. Maybe we'll do those as the like, we'll have like a paid only version of the podcast that only CTOs get to hear or some other people have suggestions to have. Hey, let's have CEOs on and let's ask them when it comes to it, what's your single biggest frustration, problem or concern? So we can hear from their point of view too, because it's really not about us versus them. It's about technology as a business force. Multiple player winning the day. So, Tim McCrossen, you've been heard.
• 1:05:00
  Speaker A: Thank you, sir.