What's Up with Tech?
Tech Transformation with Evan Kirstel: A podcast exploring the latest trends and innovations in the tech industry, and how businesses can leverage them for growth, diving into the world of B2B, discussing strategies, trends, and sharing insights from industry leaders!
With over three decades in telecom and IT, I've mastered the art of transforming social media into a dynamic platform for audience engagement, community building, and establishing thought leadership. My approach isn't about personal brand promotion but about delivering educational and informative content to cultivate a sustainable, long-term business presence. I am the leading content creator in areas like Enterprise AI, UCaaS, CPaaS, CCaaS, Cloud, Telecom, 5G and more!
What's Up with Tech?
Meet Your New Coworker: The AI Agent
Interested in being a guest? Email us at admin@evankirstel.com
The world of artificial intelligence is evolving at breakneck speed, and nowhere is this more evident than in the emergence of agentic AI - autonomous systems capable of taking independent action on our behalf. From booking travel arrangements to conducting complex research that would take teams of humans weeks to complete, these technologies promise to transform how we work and live.
Dave Nicholson, CTO advisor and industry analyst at Futurum Research, brings decades of technology leadership experience to unpack this trend. Having served in executive roles at EMC, Oracle, and Dell before joining the analyst world, Dave offers a uniquely practical perspective on AI's real-world implementation. His current work teaching in the CTO accreditation and AI programs at Wharton puts him at the intersection of academic research and business application.
The conversation traces AI's evolution from simple prompt-response mechanisms to research-capable systems and now to truly autonomous agents. While the technology's potential is enormous, Dave cautions that most current implementations labeled "agentic" are actually sophisticated versions of robotic process automation rather than truly autonomous systems. The real winners in this space will be organizations with well-organized data and the ability to extract business value from rapidly evolving technology.
For business leaders looking to capitalize on agentic AI, Dave recommends a two-pronged approach: personal experimentation with AI tools to understand their capabilities, combined with systematic efforts to prepare organizational data for AI applications. The executives who dismiss AI as something they "don't need to understand" risk falling behind much like those who once claimed they would never need email.
Perhaps most profound are the questions agentic AI raises about work and society. While these technologies promise unprecedented productivity gains, we have yet to determine how those benefits will be distributed. Will employees who become dramatically more productive through AI assistance receive better compensation, more time off, or simply face higher expectations? As Dave puts it, "We have a tiger by the tail," and anyone claiming to have all the answers about how this will unfold is being disingenuous.
Ready to explore how agentic AI might transform your organization? Start experimenting with these tools today while preparing your data infrastructure for the autonomous future that's rapidly approaching.
Discover how technology is reshaping our lives and livelihoods.
Listen on: Apple Podcasts Spotify
More at https://linktr.ee/EvanKirstel
Hey everybody, really excited to dive into the world of agentic AI today, the buzzword of 2025, with a true industry expert and insider from Futurum Research. Dave, how are you?
Speaker 2:Good, how are you Evan?
Speaker 1:Good Thanks for being here. Really admire your work and the team's work at Futurum, but let's dive right in, because it's not just about AI and automation anymore. Oh, these are top of mind. It's about intelligent decision-making at scale and agentic AI. Before that maybe set the scene here, introduce yourself your work as an industry analyst and CTO and a little bit about your background at Bio and CTO, and a little bit about your background at Bio.
Speaker 2:Yeah, yeah, so I won't start with the paper route that I was an assistant on at age eight and then took over full-time at age 10, which I'll take part of it. But no, in the 90s I started in the data storage business, working for a series of startups where we built things that competed with the likes of NetApp and EMC. Long story short, ended up at EMC for 16 years, finished my tenure there as chief strategist or chief of strategy for emerging technologies. From there went to Oracle where I was the CTO for the cloud business group. From there went to a little division of Dell where I was the CTO for the cloud business group. From there I went to a little division of Dell where we built the Azure VMware solution. And then after that I kind of went off on my own and partnered with folks in the media and analyst space.
Speaker 2:So I'm not a traditional analyst 30 years of more kind of CTO type experience in the field and then doing private consulting and when I say consulting I mean me and some private equity folks or venture capital folks helping them connect the dots between technology and business value very much in private.
Speaker 2:So didn't blog, didn't tweet, not a lot of stuff out on LinkedIn and then along the way was introduced to some folks in academia. I did a guest spot for the CTO program at Berkeley, where I was introduced to an organization that manages the programs at Wharton. Lo and behold, I've now for four years been an instructor in the CTO accreditation program at Wharton and I helped co-found the AI program at Wharton. The professor that runs the program, professor Sonny Tambay, runs the program and I'm his Robin to his Batman sort of the practical experience sort of guy. So I've got my feet on the boat and on the dock. And on the Futurum side of things, we are kind of a full spectrum analysis, advisory and amplification firm. We do deep research with our research platform. We have analysts that do the traditional analyst work and then we have this amplification vehicle where we do a lot of media. So I know that was a bit long winded. I'm going to take a sip of coffee after that.
Speaker 1:No, I had no idea. Your background is pretty fascinating. So, on a jet to go. When do you think you first heard this as a heard this as a term of art? I was thinking about the other day and I'm thinking, gosh, maybe a year ago, but maybe it was around before then. What was your aha moment when it started to come to your fore?
Speaker 2:So, um, so my friends know that I'm. I will freely admit my ignorance on subjects, because I'm confident that I do know some things about other things. Uh yeah, it was about a year, maybe a year and a half ago, and the first time I heard the term was when I was being offered a CEO gig of a startup in Agentic AI, and I didn't know. It was the first time I had heard the word agentic. I had to get them to slow down so that I understood that it was oh, agentic, okay, oh, agents.
Speaker 2:Okay, you're not talking about agents running on a server. Oh, you're talking about AI, not just saying things or doing things or saying things, but actually doing things. Okay, got it. And from there, all of a sudden, it's like when you buy a new car and you see that same thing of car everywhere, all of a sudden, all of a sudden, it's like when you buy a new car and you see that car everywhere, all of a sudden, all of a sudden, I started hearing about it everywhere. So, yeah, it's, it's come on very quickly, fast and furiously, to say the least.
Speaker 1:Yeah, no, it's amazing, and I actually was racking my brain and realizing wait a minute. In the in the nineties, in the late nins, I heard some internet researchers talking about intelligent agents, and they were even doing demos of intelligent agents that would go out on the internet. I had no idea what they were talking about. It seemed like science fiction. These were like PhD-level researchers, but imagine that that probably has deeper roots than we know. And how do you define agentic AI, given all the buzzwords and acronyms floating around?
Speaker 2:Yeah, yeah, yeah. So, first of all, it doesn't have anything to do with wait. You have to run an agent on every one of my servers. No, I have people still getting confused about that.
Speaker 2:The way that I talk about it is, you know, in sort of phases. You know, phase one of LLMs was this idea of prompting and getting a response, near instantaneous response. Wow, it's like I'm talking to a really, really smart person. The next step was more deep research. It's like, hey, don't just give me a response right away, but I want you to go away and really, really think about this and, in fact, don't just think about it, go out and research it. Go out and retrieve information and synthesize it for me.
Speaker 2:Phase three is this idea of agentic. It's where this agent working on your behalf as an assistant can go out and do things for you as an assistant, can go out and do things for you, whether it's book me airline tickets or go out and do deep analysis that I would hire paralegals to do. That would otherwise take a week and 10 people. Now, this autonomous synthetic being essentially is going out and doing this in sometimes minutes, but it does. You can segregate it in terms of time frames. Phase one prompt response instantaneous. Phase two prompt go out and think for minutes, 20 minutes, half an hour. Come back Versus true agentic is these things go out and they do things autonomously forever, right.
Speaker 1:Yeah, no, it's an amazing concept. I'm still struggling for real world examples of any of this where it's making an impact, but maybe I'm being too hard on the teams out there. Where are we with real world deployment versus the science projects that are ongoing?
Speaker 2:Well, I can tell you that I have a bit of an advantage of having a front row seat to this. I'm literally grading the papers of the leaders in IT who are in the Wharton program, seeing what they're doing, and I can say that I share your measured approach to this. Let's say I didn't say skepticism. The point is a lot of AI that's being deployed right now. The majority of it is still what we would call machine learning or deep learning. Neural networks block and tackle stuff. Very little of it or the leading edge of it, let's just say, is generative. So we're still struggling largely to get real ROI out of agentic AI at the business level. Now I have seen good use cases for that, for generative AI being used specifically sort of the ingestion of your proprietary data in a way that then allows your bespoke model to reason over it. A lot of really clever things are happening there, but we're in sort of the beginning stages of that.
Speaker 2:On the agentic side. Frankly, evan, everything I've seen that people have called agentic has been RPA, robotic process automation. It's been a little bit slicker version of stuff that you and I have known about for a long time, so a lot of it is just kind of block and tackle automation, and there's nothing wrong with that. But I think our industry suffers from a massive fear of missing out, not just on the customer side of things, but on the behemoth side of things. It's a clash of titans. Look at AWS's announcements last week. People cannot. They're climbing over themselves to invest hundreds of billions of dollars in this to be at the leading edge. So I definitely have. I might have a couple of good use case stories, but I agree with you, we're at the beginning stages.
Speaker 1:Good to know. We're not too far behind, then, and let's talk about winners, maybe not losers, we'll just talk about winners. But are there certain both companies you see as potential big winners here, or even maybe broadly speaking industries big winners here, or even maybe broadly speaking industries, verticals that could be real winners as this gets deployed, I think, as it inevitably will, some fashion?
Speaker 2:Yeah, yeah, I would say that the sort of horizontal that we have to look at in terms of industries not the suppliers in this game, but the actual consumers to the extent that you possess information, you have a chance to win. And that goes to the answer to the question that you didn't ask yet, which is well, where do you start? Well, get your data in order right. So there's a lot of basic organizing your infrastructure that has to happen, so it's AI ready. So the horizontal where people win on the consumer side people with data, anyone with data you have a huge opportunity. On the sort of supplier side, I think a lot of the incumbents are going to do very, very well what we would call the hyperscale cloud providers. They're going to do very, very well In the consumer space. I think that the kind of Android, apple universes universe I I don't know, is that a term that they will both do well. I think Apple is going to do well, despite people's criticism about their laggard-ish behavior.
Speaker 2:I think that what we would loosely term global systems integrators, services organizations, consulting organizations they are going to be the key to a lot of this, because the Googles, the Amazons, the Microsofts of the world with their LLM with their model partners. Notice I didn't mention OpenAI or Anthropic or any of those. I believe that they will fade into the background and become features of the environments that businesses use. So if you're using AWS for cloud services, agentic AI, generative AI, will be a part of that. You won't care what the name of the screw is, that the bolt is screwed onto at some point.
Speaker 2:So I think those things are going to fade into the background. But the consulting folks are going to be the key to this, because Amazon, microsoft can come up with the shiny toys. Getting business value out of those shiny toys. It's all where the rubber meets the road. It goes back to kind of the MBA admonition of first understand the business problem you're trying to solve, first understand the opportunity you're trying to pursue, then look at the tooling that becomes critically important. So I think the big winners are going to be folks who understand how to extract business value from technology that is moving at a pace that is unprecedented. They'll be the big winners.
Speaker 1:Love it and you know we in tech we love our shiny, flashy demos and apps. We love to talk about all the cool stuff we're using. But maybe you're a tech leader or a CEO of a great but boring business. I don't know. Trucking or warehouse distribution. I mean, where do they start? What's a good way to experiment with something like this when you have a real business to run?
Speaker 2:Yeah Well, so remember, I alluded to that MBA admonition of first seek the business problem. I will now completely contradict myself. I think that in the age of AI, it's absolutely critical that you first of all don't try to learn this stuff. You can't learn it, nobody can learn it. Nobody can learn it, nobody can learn it. Maybe you can. I can't, I'm not smart enough. But I'll tell you what I can do is I can play with it and everyone can play with it.
Speaker 2:So I would say to the CIO or the CEO you need to start playing with these tools. Pick two or three, start playing with them. And I have folks like this who will tell me oh, I'm the CEO, I don't need to know the technology. And I will look them in the eye and say you know? You remind me of the guy who said he would never have email Because I'm old enough to remember that. And those guys got ran over by the bus. You'll get run over by the bus if you don't have a fundamental understanding. And I back that up by giving an example.
Speaker 2:If I said, evan, here's a thousand acres of land, what are you going to do with it? It's yours. First, you say thank you, dave, for the thousand acres of land, plant a tree. Well, wait, wait, hold on, hold on, because you've only seen shovels. Before you answer, evan, let me show you a little video clip and I show you a. I think it's a D9, earth mover plowing a road. Just 10 second clip, no sound. First you'd be terrified because you wouldn't know if it was a dragon or a machine or whatever. It doesn't matter. You've now completely had your imagination opened up in terms of what's possible.
Speaker 2:So I think it's critical that business leaders themselves play with these tools. You need to be able to understand how to prompt a generative model and get what you want to create an image, to create whatever it is, to do some research, just to have a fundamental understanding of it, so that your imagination is broadened. Then the first step goes back to, at the same time, get your data house in order. If your data is siloed and it is not ready to be accessed, to be reasoned over by generative AI, you can't benefit from it, and there are a lot of decisions that have to be made in that regard. Do you allow your data to be in a cloud somewhere or are you going to do this on-premises? It starts going back to the fundamental questions of infrastructure. So the two things play with the tools. Don't wait, start playing with tools now, the individual contributor type tools. And the other thing in parallel is get your data in order so that you can leverage this stuff moving forward.
Speaker 1:Fantastic, and so the promise of Agentic is truly autonomous operation, which is exciting but kind of daunting. We can't even manage people these days. How do you manage agents? What do you think that means in terms of people and culture and just the way we work? How is that going to unfold?
Speaker 2:Yeah, so is it fair to assume that you have not signed over power of attorney to an agent yet?
Speaker 1:It's fair to assume that's right.
Speaker 2:Because neither have I and I'm not planning to in the near future. We don't know. We don't know what this is going to look like in the future. I can construct an ideal scenario, but even the most ideal scenario has challenges. The ideal scenario is that you and I, as enlightened individuals, learn how to leverage these tools to make us massively more productive, whatever that means. However it's managed in our work, in our personal lives. So we become much more powerful and we reap the benefits from that in whatever way we feel is positive Make more money, have more time off, have more time with our families, whatever that is the point is, in the ideal scenario, having these agents working for us means that we're more productive and we reap, as individuals, some of those benefits.
Speaker 2:There's no clear design for that yet. If an employer puts these tools in your hands and you can now be 300% more productive, does that mean that you get an extra Friday off every month? Does that mean you get 20% more pay, or is it well? No, if you don't keep up, we'll find someone who can keep up, so we're no closer. Widen the gap between those who know how to use these tools and those who do not, and that is going to cause societal upheaval. This is not a technical conversation at this point. It's going to cause problems. The days of a single wage earner being able to work on a factory assembly line and support a family are long gone. But when you look at something like trucking, which I think is the highest paid non-college degree requiring gig, that's out there.
Speaker 1:And the biggest occupation in the US.
Speaker 2:Yeah, you talk about autonomy entering that space. Well, guess what? There will be some people who will sit at mission control and they will control 30 trucks like a drone operator would, and they're going to do very, very well, and they're probably going to make five times as much money as they did when they were driving the trucks, and they're going to be at home every night for dinner. But for every one of them there will be 20 or 30 others that will be massively displaced. So the short answer is we don't know how this is going to unfold. We do have a tiger by the tail here, and anyone who says they've got it figured out is either dishonest or at least disingenuous, or at least disingenuous.
Speaker 1:Interesting Times ahead for sure. So, you know, in terms of, you know, implementation, obviously, experimentation you mentioned. Are there any playbooks or rule books or books that you recommend to be old fashioned for leaders looking to journey down this path? Or is it just going to be discovery?
Speaker 2:Yeah, yeah, Books. Is that what you said? B-o-o-k book? I've heard of these things Old fashioned Audiobooks.
Speaker 1:I've got books I see.
Speaker 2:Yeah, no. I think the way that I would answer that question is and I mentioned it at the outset you know, look, 30 years of being on what I would call the vendor side of things. The best case scenario when you're engaged with a vendor who's selling you things is that you get the really good objective advisors that work in all of those firms and you essentially get them to do free consulting for you Because you can have a company and you essentially get them to do free consulting for you Because you can have a company. I'm going to pull one out of thin air. You could have a Dell Technologies come in and essentially give you millions of dollars worth of free work. People at Dell would hate me for saying that, but it's true. Yeah, okay, maybe you're eventually going to go to McKinsey or Accenture or whatever for some deployment thing, but in terms of figuring out how to connect things together and where to start, the playbooks that exist, that are the closest thing to leading edge, exist in the hands of vendors. It's the Dells, the Oracles, the HPs of the world. They have it and they're doing it internally. Um, I, uh.
Speaker 2:I had a chance to have dinner with a guy from hpe at hpe discover a few weeks ago and um, uh, I don't want to share too much, but, but essentially it was this idea of first they got their data in order and then they reasoned over it and they were able to figure things out that would have taken them years in the past, and these are things you'll never hear about because they're not very sexy. When you say, why is it that we're having trouble with these kinds of contracts, and the answer comes back immediately it's oh well, you thought you had three types, you have seven. You have seven versions of these things lying around. And so the vendors, who will live and die based on the success of their AI implementations. They do on behalf of their customers, they're doing these things internally and they have the most to offer.
Speaker 2:I would put a large hyperscale cloud provider or an infrastructure provider. I would put their experts up against any experts from an Accenture, a McKinsey or a Bain, any day of the week, any day of the week. So leverage your vendors. Understand they're trying to sell things to you, but you should. If you're the person who has always said to a vendor don't refer to us as your partner, you are not our partner, we are your customer, you are our vendor. You're going to get run over by a bus and I will, and I will dance a jig at your funeral because because that is the dumbest thing ever you need to leverage these people who are forging the path right now.
Speaker 1:Yeah, the stories are kind of dripping out week to week, month to month. They're pretty impressive. I mean, you hear about Salesforce not hiring engineers this year and the agentic revolution happening internally and the agentic revolution happening internally. You talk about IBM laying off sort of legacy workers and functions like HR and hiring people.
Speaker 2:In AI it's everywhere. Yeah, all those examples, I know we need to treat those with a fair amount of skepticism because, look, you can be a big company that wants to increase profit by reducing cost, and what's the best way to do that? Headcount right. So it's very easy to use we're using AI as an excuse to reduce headcount. The difference is when they prove it to you, if they can show you, if they say you know, by the way, we have 90% of the people we did a year ago and we're just fine and it's because of AI. Well, they have just put a huge burden of proof on themselves. They need to show you exactly. Well, what is it that you're doing that allowed you to eliminate these people? Somebody just is completely outsourcing marketing. It was Intel, right? Intel announced they're completely outsourcing some aspect of marketing and the people they're outsourcing it to are going to do it with AI. Okay, well, at least they've presented sound reasoning. They're not just saying they're going to use AI and lay off people. They've got some proof behind it, so demand proof.
Speaker 1:And clearly. Just on a final note, there seems to be something big happening in software development like a tidal change, sea change in how developers are working and with what tools and how code is tested and automated and even developed, and there seems to be evidence that Gen AI at least if not agentic yet is having a huge impact on the amount of work that developers can do, how much they can do, and the expectations now are that developers have to be 20 or 30% more efficient with these tools they are now given. Is this the canary in the coal mine for the rest of us who are not software developers?
Speaker 2:Yeah, I think it is. I think it is, and I talked to a lot of people about this and there are some that are adamant about pushing back against this.
Speaker 2:It's like, oh, it's not as good as people say it is. And then I talked to people who were like it's better and it's better and it's scary. And I think what it comes down to is you know, it's kind of like the 80-20 rule. I think 80% of what human developers do can be replicated adequately by these tools. And that's the challenge, the leading edge no, and that's the challenge, the leading edge no, I think that that you know there's a thing called the vanishing point on the highway, where the two sides of the road come together in the distance and you never actually reach that point, because as you get closer, it just gets further away. I think we're always going to be in that situation with this stuff. There will always be a leading edge where humans need to be in the loop, to use the phrase. But I think that it is. Yes, it is very, very real. It's hard for me to directly evaluate some of these things, because when.
Speaker 2:I play with these tools. When I vibe code, the code that's generated might as well be Klingon to me, because I don't have the background writing code, so I can't tell you whether it is accurate or elegant or not. All I know is that if the code generated, when executed, creates the dancing monkey gif that I asked it to create, well then I consider it a success. What I'm not clear about is when you have all of these LLM generated code things, how well will they work when they need to work together? So one example I saw recently was with some ServiceNow workflows, and this person was absolutely could not wait to show me. It's like look, I have no background in this and look what I built and in a vacuum. It looked amazing.
Speaker 2:But the question is okay, but how will that interact? Will that break other things? I don't know. I don't know. So beginning stages. But yeah, I think it's going to be massively disruptive and the ones who know how to leverage the tools will benefit. The ones that don't will fall behind. It's going to be interesting times ahead.
Speaker 1:Interesting. Indeed, I can't wait to have this conversation in a year from now, and we're going to have way more data points and insights and analysis from you and the team at Futurum. Thanks for joining. Really appreciate the insights.
Speaker 2:Thanks, evan, great to see you.
Speaker 1:And thanks everyone for listening, watching, sharing. Always appreciated this episode and be sure to follow our new TV show now on Bloomberg and Fox Business at techimpacttv. Thanks everybody.