Why Enterprise AI Fails — And the Skill That Makes You Unfireable

the AI title wave flip- dogong and someone who usually is like a seale person in the company was given this responsibility and they went and got tools the most common things and they just handed it out to employees and they started using stuff the governance and security of company data wasn't happening what they were using it for how was it making their lives better how was it returning an ROI not happening will it replace everyone's job I think that. All right, guys. Well, whether you're an engineer, an architect, or in management in tech, you probably thought about starting your own tech consulting firm, but you probably had no idea how to do it. So, today we're going to do a deep dive into that space with a guy who has done this for 22 years. His name is Evan Solomon. He has EFS solutions, and he is both an AWS partner and a Service Now partner. He's actually one of 65 companies to reach the highest tier of AI consulting in AWS. And Evan, welcome to the show. Thank you. Yeah. So, like I said, a lot of people that are going to be tuning in, they've thought about doing this. They have a highle job in tech. They thought about making the leap. They see some of the bills being paid to companies like yours and they say, "Wow, I'd love to do that, but how do I get started?" So, how did you make that decision to get into this? Well, I went to Drexel University. that was very pre-professional and we had this co-op program where you were working as you were studying and I had some really amazing experiences in New York City and Viacom and Black Rockck Financial and I saw that the IT department was just totally underappreciated. I had a mentor who told me once that you know it is only noticed when the status quo is wrong and that value proposition and I said to myself you know I'm never going to get my beautiful mind value out of that you know I knew I had some special skills I knew I had this wherewithal and I just took a plunge and at first it was being a consultant it wasn't a company like it is today, 22 years later, but damn, the money was great. You know, whatever my initial job offer was back then, I think I had something at Goldman Sachs. I was making hourly wages of $75 to $95 an hour, you know, circa 2005. What was that role? It was a management information MIS which is a word they don't even really use anymore but it it's like the business and IT fusion and just running ops. Okay. So you eventually start as an hourly consultant through your own company your own firm that that's what EMS started. What point because I know a lot of people that are you know they have their own independent corp. How do you go about getting your first employee under you and also selling the solution services that you do now? So, it was stacking up, you know, long-term value in a customer and also bigger projects and finally making those difficult decisions as a sole proprietor that's like, I can't do everything. I'm going to have to take, you know, some off the top here and work with others and it's going to ultimately let me grow. And I would say that for years I battled that in the beginning. Okay. What was the the back and forth? You saw too many opportunities that you could get your arms around. So you said, "Wait, I have to build out a structure where I could kind of get more of these while I'm overseeing the the top line." Yeah, absolutely. And I had a great ability to describe technical solutions to business problems, but I realized at the beginning that it was just too much. It was too in-depth. You start with, you know, a small business into the smaller mid-market company. It's hard for a business owner to understand all these things and they're scared about technology. what's it going to do against the status quo is already hard enough to understand for them. Let alone where the value prop and the strategic stuff is coming. And back then that was a lot harder to describe. Now there's so many more use cases and there's so many more incredible things that have happened that even the small business owner understands that value prop more. But that really forced me into understanding how to do like this translation language, how to pick the things that are the most important and deliver the explanation around that and get the business owner, you know, behind it. And not only did I do that in the way that I describe these things and sold these things on the proposal, I actually did it in the deliverables. So trying to sell a project that's a year or two years long that has 27 phases like their head's just going to fall off. So we even to this day even with our significant growth we sell things in compartmentalized phases and we also sell things that are doing things and still learning and scoping things instead of just being like a big consultancy that wants to sell you a giant scoping project for two years to figure out what we might do or there's a freelancer who's ready to make a couple bucks and he wants to start today. We sit in between the two of those. So, we obviously need to still do some foundational scoping, but we get started with some kind of phase 1 project could take 6 weeks. And in that time, a lot of really important things happen that are mutually beneficial cuz we're basically dating the customer in that time. So, we're we're learning as we go. We're building trust in each other as we go. We understand. You may tell us that this is the most important thing, but as we work together, we some other rabbit comes out of a hat. So after that first phase one, the customer sees that we really care about what we're doing. We're responsible to our goals and deadlines and communicative through that whole process. And we deliver something of value, not something that's the kitchen sink. And then they're like, "Okay, you're the easy button now. What do we do next?" So get in at a, you know, kind of a micro level and then they say, "Wow, this guy delivered." And then you kind of come in and say, "Well, we actually also do could do this for you." And you kind of upsell once you kind of get in. Yeah. But that's another thing, the word upsell and the whole business of selling technology and this was really by accident because we were small and we didn't have dedicated salespeople but it's something now that I will keep forever and that's where you have the practical knowledge and the practical experience and the war battle wounds of stuff messing up that speaks trust. And when you put that into your sales, that's what really I we've seen, you know, connect with customers on that. So all of our sales folks are engineers. They've been in the trenches. They've done stuff. And when they make a promise to a customer, they can foresee how long it would take them to do it themselves before, you know, the sales guy is like, "Hey guys, we just sold this amazing thing, but we have no idea how to do it." and we have nobody who can do it. And I told him it was going to take a week, but it's but it's not going to take a week, right? So that genuine part of ourselves that really like has this flare of being like sole proprietors and freelancers and just a bunch of guys in the trenches and then the other side where you're like this puffy company that has 19 different departments that has to answer your question. Finally, we sit in the middle. So, we have capacity. We have a baseball team. We have a deck of guys that are ready. I call my family. So, we get that component of a larger company, but we still have that grit and that roll up your sleeves and that care from the individual guys. Yeah. And you said you started in 2004, right? Obviously, I mentioned you were AWS partner, Service Now partner. That wasn't a thing back then. So you eventually acquired those partnerships. How did that happen? And like kind of like take us through what quickly what's the process of doing that? Absolutely. And that's another thing as a smaller IT firm or an individual guy. I had a mentor that taught me hitch your wagon to somebody who can pull you up. So and sometimes you know you can hit your wagon and get lost too. But we've had probably over 50 different partnerships in 22 years and we've narrowed it down to the ones that are the kinds of customers that get our value prop and the kinds of customers that are sticky and we AWS we saw that journey of onrem back in the day into the cloud and we saw that company just go hog wild and we also saw so many different tools and services that they offer and that the way that they operate is in a is in an innovation forward and our guys were drinking the Kool-Aid in the training and the skill development. So we hopped in and became a partner. Like any good partnership, it has to be mutually beneficial. Like you can't just sign up with AWS Microsoft Service Now and expect to be like, okay, here's millions of dollars of contracts, here's leads, you know, it has to have go both ways. So, we've spent a lot of time in AWS University, if you will, as individuals. We've spent a lot of time in AWS University as a company getting some of these very coveted like this AI certification that only 65 of all of the AWS partners in the world have. And how many partners are there? over 250,000. Wow. Jeez. So, we've we've put in for AWS and in return now we're seeing the dividends and and that symbiotic relationship has really been amazing. Um, and there's one really more interesting important point which is that these companies the way that they measure their revenue is all in SAS. So, they want monthly fees. They don't want the services business. And that's where this value added partner relationship is the symbiotic thing with these bigger partner relationships. So they're trying to get the monthly number, but all the steps that you need to take for that company to be spending that monthly number to be maintaining it to make sure it's secure and reliable, Amazon doesn't want any part in that. AWS doesn't do that. And that's where they come to the partners. And the partners act as kind of the sales force for them and maybe you know kind of like doing that upsell right yeah they I would say that the partners can put a personal touch on you know here's this giant company that has incredible capabilities incredible services some you know that no no one else can even compare to some of the services they offer but here's the personal touch. Here's where like I'm going to take you as not a number and as understanding your business journey, the specialty things that make your company unique and how am I going to wrap that around the AWS cloud. Yeah. And for them to serve because AWS is pretty much in every company for them to service them with their own people, they would have to hire I mean god knows how many more employees and that would be a nightmare for them. Correct. Yeah. I mean, that's pretty much the business model of SAS companies in general, hyperscalers, any enterprise software company. Like that's that's just the industry standard. Like they don't do the implementations or they only do like certain implementations. They almost price themselves out of the implementation. Sometimes the hourly rates for hiring them directly like they almost purposely put it in such a tier that's like right only the very few who want to do it. If you want to pay it great but you could just go to these guys and yeah so then so in the partner network you know everyone is given those opportunities because the main enterprise service that you're a partner of isn't doing it. But then you still got to set yourself apart and that's where finding the niches is really valuable. Yeah. So you're who are you kind of competing with other firms on your size that also are targeting these kind of niches in your partnerships? So in 22 years we're not a small business anymore. We're not competing with individuals and small business consultancies, you know, with less than 10 employees. We're a mid-market company. Mhm. Uh so we're competing with other mid-market advanced tier AWS and Service Now partners and we're also competing with you know with the big consultancies of the world the KPMGs Price Waterhouse Coopers Capgeeminis etc. And you know it's hard as we get into the major enterprise customers that to break that nut from the consultancies. Um, but I think that there's some really novel and interesting things that we've created that may even be con complimentary to the consultancies and you know we don't have to push these behemoths aside. we can all play along together. And there's been a couple deals that we've done as subcontractors with the giant consultancies. And I mean, there's plenty to go around. If if you can build trust and you really can deliver on what you're talking about and you can really get that ROI, then there's plenty to go around. What about obviously we mentioned you're one of, you know, those elite AI partners for AWS now. So you get into these companies, they're trying to stand up their AI thing and it flops. Then you come in correct. So why is it mainly flop for some of these big companies and how do you Yeah, I would love to talk about that and I would love to talk about like the AI journey that we just see the same story over and over again and we see it in mid-market companies and we see it in major enterprises. So the AI title wave flipped on and someone who usually is like a seale person in the company was given this responsibility give AI stuff to people in the company and they went and got tools the most common things like a Microsoft 365 co-pilot or whatever it may be chat GPT and they started using stuff. the governance and security of a company data wasn't happening. What they were using it for, how was it making their lives better, how was it returning an ROI? Not happening. But that was phase one. And a lot of companies are still paying for those subscriptions. You know, 50 bucks a month times however many employees for co-pilot. Then the structure came in and there was either consultancies or some of the larger tools that started to build MCPs and collect data and get some kind of a structured component that has some kind of foundations of governance and security around it. And that with that phase two, I would say that most companies in the larger enterprise are fully into that phase two and mid-markets are in the more beginnings of that phase two. So then we get to phase three where okay, you're using AI, it's now more integrated into the organization and it's serving a purpose, but what purpose and at what benefit? At what ROI? Okay. And that's where we're seeing this screeching halt right now where they're looking at, okay, co $50 a month times 5,000 co-pilot license, a millions of credits on the LLM, and where is it going? How do you calculate the ROI? And sometimes that wasn't even considered in the beginning of the project at all. So we see this tier three problems about ROI and then tier 2 problems about even getting a foundational data structure that you can use AI with and that problem is also a whole another conversation. So this foundational data structure and the business process stuff that is something that's always been in existence and say you're a small business and suddenly you get exponential growth and you become a mid-market company and then you get private equity funding and you acquire other branches, franchises, whatever it may be and then you become a small enterprise. It's a rat race. It's a hot mess and it's hard for a small business to be keeping up their business processes with that growth. So we find that a lot of companies don't have well doumented business processes and that was already in existence. So if you're going to take AI and you're going to automate those business processes or streamline them or minimize those business processes, then you need to know what the business processes are. You need to have them as well defined and who's responsible for them and who what are the inputs and what are the outputs that you expect because you can unleash Gen AI just asking lots of things about your business and then you can unleash a Gentic AI where it's like an autonomous bot that's doing work on its own but it can just create more chaos if we don't have a foundational data structure. So the old words before AI, machine learning and even before that just data science those skill sets are absolutely incredibly important now because it's this this is what needs to happen that data has to be recognized where it lives how does the business operate why before you can just bring AI into the And we've been through that battle many times with customers. And let me be honest, another thing that customers build our trust with. Sometimes we've failed miserably and we're not afraid to admit it. And from those failures, we've actually found interesting things under a rock or we've been able to teach our employees and our customers how not to navigate into those solutions. And in 22 years that has really been an amazing journey. What's like what's an interesting thing you found under a rock at a company that I mean I guess generally speaking I don't know if there's NDAs or whatever involved but what's in a general sense what's something you found that they were like oh wow we had no idea about any of this. So I would say that it has two columns cost where someone's paying the bill on a service like AWS and maybe they look at the top level subcategories like how many EC2 servers or how much bedrock is using per month but in the details the devil's in the details and they the cost impact of certain things have definitely been uh found under Iraq. And sometimes they're not supporting like critical infrastructure or critical processes and the relationship between kind of throwing money at the problem like whether that's in the software or whether that's in the business process you know only gets you so far. So that's one half is the cost component. The other is the technical and security component where we have found just basically gaping holes in software infrastructure in cloud infrastructure and whether that's technical flaws or whether that's opportunities for human engineering because that's some of the worst hacking that happens now is not just the the the SQL injection is an opportunity and discovering ing those gaping holes have been interesting for sure. Yeah. So, obviously you have kind of a unique view of how these companies are attempting to use AI now. Seems to be a pullback. I I know we mentioned before the episode Uber said they already spent their whole AI budget because people just ran through the credits. Um, where do you see this going? Will it replace everyone's job? Are these companies going to pull back and just kind of regroup and see how they can deploy AI? AI better like what's going to happen in the next six months to a year? I don't particularly seeing a pullback. I think that boards are still really really pressing hard on organizations for ROI from AI. And I think that they're going to have to skin the cat differently and they're going to have to cut some tools or cut some other processes that aren't delivering. But I still see that gas pedal down from boards and from the top because there's definitely still the the strong smell of of savings. Mhm. Now, what was the other part of the question that you asked? Just where it's going and will it replace people's jobs? So, that's a great question and um I think that there's going to be displacement. There's no question about that. It's more a matter of the size of displacement and the type of displacement. So, if you are someone who I consider one of my favorite words has a license to learn is something that I built my company. I'm 22 years of our family of employees. a license to learn. It's not even like what is your specific skill set that you have x years of experience in. It's can you do research? Can you piece things together and solve a puzzle? And if you have that skill within your industry, that's very valuable. And in fact, it might even be more valuable with AI because that's where you can help answer to the board where's the ROI in this thing or which turn should we take that we need to get to the ROI. So if you have innate problem solving skills and you have intermediate knowledge in your industry, I think for now you're pretty safe. Now, how do you I mean obviously you're an expert in the field. How do you think people should navigate things going forward? Because like let's look at the last two, three months, right? Clawbot was taking over two months ago. Every like you were a loser if you weren't using Clawbot. Then everyone was on claw code. Then they ran through their credits. Now the chatbt agentic AI is like making a little comeback. So, how do people kind of move forward and kind of push out the noise while trying to kind of obviously work on their careers or businesses? Sure. So, the hottest thing is AI discussion and everything that you can read. And I would split this up into two categories that I don't think people are splitting up. There's consumer AI and there's what is a business going to use AI for that has a strategic value and ROI. So a lot of the hype I think is consumer AI and and this happens with any new technology when it's so new it usually starts out as a consumer AI because it's no big deal if your iPhone crashes but it's a big deal if a company loses data. So, and then that consumer AI starts getting used by businesses without it even being properly flavored for business use. And that's what I think we've seen happen in AI. And the word that I've heard one too many times, governance, which is checking just like a human, what access does the AI have? How much could it do for good or for bad with that access? Where's the kill switch? How do we keep track of what it's doing? And how do we do it in a way where it's just not log hell that no one's going to even know, you know, until we spend two days in the forensic analysis where we can actually get actionable things. So I think tools that are this mid layer between the LLM and the actual business processes that create this governance are really really important. And as the fads and trends happen with whatever LLM is the greatest thing, these tools that sit in the middle are what are going to be more sticky and more valuable. What are some of those tools? Okay. So, some of the tools, well, Service Now has AI co-pilot, which is a plethora of tools that help with the ingestion of data and getting it into the right place. We've worked with another company called C Data who has a package of MCPs and an MCP is basically like the API that connects outside data. But the more interesting thing in these tools, it's not just about whether the the bot has access or not. It's also about cost because you said Uber spent all the credits, So, it's not just governing the security, it's governing the credits and what they're used for and why. And if you can govern the credits, you're also building foundation for ROI because you could see this thing used X credits and it produced Y results. So it's also it's a multirong. Yeah. So you see as far as in the age of AI, if that's what we're going to call this, right, governance is not going to be outsourced to nonhumans. Correct. At least right now. Someone's going to have to run those tools you were talking about, uh, that layer in between. Correct. And and they're going to keep that a human, right? Yeah. I mean, it's impossible for a human to crunch all the numbers of everything that's happening in an organization. AI can do that, but then the human has to be involved in understanding and building the business processes around the governance and where to be looking at the problem. like AWS had an outage uh last year in the third quarter and it was caused by AI out of control and just the timing of being able to hit the kill switch of a human being knowing that the escalation has happened and we need to hit the kill switch is so important before we're hours days later. Yeah. So governance, what else? like some of these analysis type jobs that can use those crunch numbers and say, "Okay, it's showing this, but maybe if we dive deeper into those numbers, it's not what it looks like." Yeah, data science is was this boring math job that now is a sexy interesting thing. And you can put a fancier AI name on it on your resume and that's what you can help with. Sure. But as far as the foundational skills of data science, there's not enough to go around because we have to have organized data. Otherwise, garbage in is garbage out. And same goes for AI bots. Yeah. Yeah. Well, why don't we uh why don't we take it home? I want you to kind of talk to the people that are considering starting their own consulting firm or they want to and what can they kind of look out for in their career right now like a niche that they should focus on while they're working their job. I would say that any of the degrees that you can get from AWS or Google or Azour that's focused on data science or data governance I would say is a great starting point. Okay. And then anyone that could use your services you know as far as a a company that has a problem who should reach out to you. So we have this incredible AI program that we are one of 65 companies of all of the 250,000 AWS program that we were selected from where we are getting cash from AWS not credits to fund any AI project. It starts at 15 to 30,000 that AWS will pay for the consulting and scoping and then whatever we learn from that and we actually do the nuts and bolts, we can get a $100,000 to over a million dollars that AWS will fund for us to do the project for you. Wow. Wow. That's incredible. Yeah. Okay. Well, you heard it there. If you are thinking about getting into this type of career, making that leap from kind of employee to consulting firm, there you go. And then obviously we'll have all of Evan's links. You can reach out to EFS Networks if there's anything that you need in AI because again he's one of 65 companies in the planet that is a top tier partner with AWS. So Evan, thanks again for making this happen. Thank you. We'll catch you guys next time.