AI Agent Builder Summit | Analyst Kickoff

Clips
News
More from AI Agent Builder Summit

Scott Hebner

Principal Analyst, AI

theCUBE Research

Paul Nashawaty

Practice Lead and Principal Analyst

theCUBE Research

Rob Strechay

Dir./Principal Analyst & Host

theCUBE Research

Trust and transformation: TheCUBE Research’s take on agentic workflows in enterprise AI

Enterprise AI is entering a new phase, where intelligent systems are built to collaborate, make decisions and achieve goals — not just generate content — with agentic workflows driving this transformation behind the scenes.These advanced workflows represent more than a technical upgrade. They signal a shift in how businesses structure automation, combining small language models, real-time data and human-in-the-loop design to create adaptive systems that deliver tangible outcomes

play_circle_outline Industry Leaders Discuss Transformative Agentic AI at AI Agent Builder Summit

play_circle_outline Focus on ecosystem of solution providers and internal contributors

play_circle_outline Importance of embedding AI into corporate culture and ensuring accountability

play_circle_outline Building Trust in AI: A Guide for Business Leaders on Transparency, Explanation, and User Intervention

play_circle_outline Overview of the "ladder to Agentic AI" including knowledge models, decision intelligence, and agents

play_circle_outline Emphasis on understanding one's own AI journey and not overspending on AI initiatives

Info
Transcript

Analyst Kickoff

Scott Hebner

Principal Analyst, AI theCUBE Research

Paul Nashawaty

Practice Lead and Principal Analyst theCUBE Research

Rob Strechay

Dir./Principal Analyst & Host theCUBE Research

HOST

Scott Hebner, Rob Strechay and Paul Nashawaty of theCUBE lead a kickoff discussion from the AI Agent Builder Summit, setting the stage for a deep dive into the business and technical implications of agentic AI. Their conversation frames the event’s agenda, which includes insights from companies such as Semaphore.ai, Deloitte and IBM.

The segment covers foundational concepts in agentic AI, from workflow orchestration to infrastructure alignment. Hebner underscores the importance of building trust in AI systems, while Strechay explores how agentic models... Read more

explore Keep Exploring

What will be discussed at the summit on building AI agents and agentic workflows? add

What is involved in creating a cohesive system that brings together various forms of artificial intelligence, expert lectures, data platforms, and technologies for effective data wrangling at an organizational level? add

What are the benefits of GenAI in terms of transforming and improving the efficiency of various tasks, such as travel planning? add

What is the importance of trust in driving adoption of innovation and Agentic AI, and what are some key strategies for building trust in the user-agent relationship? add

What steps are necessary to gain a deeper understanding and insight into the intelligence of AI, particularly in relation to domain-specific knowledge models and explainable decision intelligence? add

What are some important considerations for organizations looking to implement AI technology? add

bolt Powered by CUBE AI

Analyst Kickoff

search

Scott Hebner

>> Hello, Scott Hebner here. We are so excited that you have joined us on theCUBE for the AI Agent Builder Summit. I'm the principal analyst for AI at SiliconANGLE Media and theCUBE Research, and welcome to this industry-unique digital summit dedicated to exploring the transformative value of Agentic AI and how to build AI agents and agentic workflows. During the summit, you will hear from 20 industry pioneers from a dozen companies, sharing their real-world insights and experiences that can help you speed your way to ROI, and of course, competitive advantage. This includes Semaphore.ai, Ascendion, AgilePoint, Deloitte, SAS, IBM and Geminos.ai, and the real-world perspective of end-user businesses from Fitch Group, Emerson, Align, Hewlett-Packard and G2. Finally, our team of industry analysts are here from theCUBE Research to be on hand to provide end-end coverage and analysis, helping you put all the pieces together. In this session, we'll kick off the summit with an analyst pregame show, if you will. We'll discuss what we have learned from hundreds of market engagements and ongoing research around the rise of Agentic AI, including the strategic success factors that we recommend businesses prioritize as they start their journeys to Agentic AI, the technology roadmap for building AI agents and wiring them together into agentic workflows that are intelligent and adaptive, and why Agentic AI is the ultimate team sport in technology, all about harnessing an ecosystem of solution providers and internal contributors. With that, let me introduce two of my colleagues here at SiliconANGLE Media and theCUBE Research. Rob Strechay, our managing director and principal analyst covering data platforms, cloud infrastructure and IT observability. He's been discussing Agentic AI across a large cross-section of solution providers. And he'll help us gain a holistic view of what people are saying out there. And then Paul Nashawaty, our principal analyst specializing in application development and modernization who knows the app dev market probably as well as anyone out there and who has been spending a lot of time thinking through the intersection of an SDLC and Agentic AI. So Rob Paul, welcome. Thanks so much for being here.

Rob Strechay

>> Glad to be on board.

Paul Nashawaty

>> Thanks, Scott. This is going to be a great summit. Looking forward to it.

Scott Hebner

>> Yeah, you guys are looking really sharp compared to how I usually see you guys. Thanks for joining.

Rob Strechay

>> Got to bring the best for the audience here, especially with this highbrow Agentic AI and how people are building it. Definitely. I'm excited for this, so bring the best.

Scott Hebner

>> Yeah, there's no doubt this is dominating the market discourse these days, isn't it? Everywhere you go, everyone's talking about Agentic AI. They're talking about AI agents. And I think what we're going to try to do here is separate the signals from all the noise that's out there. I think we have a great lineup during the summit to give us some really good insights into what's going on here. All right, guys, so let's start with what you think is going to be the most interesting thing that you're going to keep an eye on. What are the learnings from the summit sessions and everything that you're going to be really keeping an eye on here? What are you looking for, in other words? Paul, you want to start there?

Paul Nashawaty

>> Absolutely, Scott. There's a lot happening. Of course, we know that AI is top of mind for many organizations. We know that AI is funded in many organizations as a separate budget for it. The impact to the application development and application modernization, AI is a big, big factor. What we see in our research, we saw... A year ago, we ran a study and we found that only 18% of production applications were running AI in the production application. We re-ran that study nine months later and found that 54% of production applications are running AI in production. Now, today, of course, we see that AI is, if you're not running AI, you had a competitive disadvantage. So Scott, there's a lot here with regards to how organizations are looking at AI, looking at AI across the CICD pipeline and across the overall STLC.

Scott Hebner

>> Yeah, I've been spending a lot of time with AI companies. Paul, you've been spending a lot of time with app dev and modernization companies. Rob, you've been really talking to a big cross-section of companies, as I mentioned before, right? From observability to data platforms to cloud. So I think you have a more holistic view. So what are you hearing about Agentic AI that's going to really from these companies that is going to make you look for something from these experts today during the summit?

Rob Strechay

>> Yeah, I think, again, similar to what Paul was talking about and building off of what Paul was talking about, because I think there's a lot of people who've gotten to POC with AI or maybe they've gotten to a first-use case. They've started out using APIs or API-driven LLMs and things of that nature, but now they're trying to get even further, and we are seeing that really a lot of the talks start to go to small-language models as we predicted about a year and a half, almost two years ago now, that the long tail with inference and small-language models, fine-tuning, very industry-specific types of AI is really where the market is going to go. But I think what people can learn out of today that it's really key is where do you get started? How do you start building on top of this? And it's not just about GenAI, it's how do you bring together chain of AI and chain of thought and things of that nature or lectures of experts and all of the different technologies that go into that. As I look at it even down the stack from that towards the data platforms, there's a lot of pieces that have to go into the data wrangling, which a lot of the organizations are going to talk about today about how they're helping with that at the agentic level and how you build into that. So I'm really excited, especially how that plays off the folks that are actually already building it. I was on with a customer this morning who's already building agents. In fact, they've released two agents this week and you start to look at, it's not easy and there's a lot of fud out there and a lot of things that people have to get through. So I think people can sit back today, watch this and watch all of the people we have on, talk to how do you succeed with this? How do you get out of death by POC? And I think that's a big key to today.

Scott Hebner

>> Yeah, we've all been through many, many cycles of innovation here over the decades. And so when you look at the Agentic AI marketplace, what's starting to occur in 2025, do you view it as a progression? Is it a transformative moment? Is it hype? How would you classify if you were to rank this with all the transformative moments that we've had over the last many decades? Paul? Or Rob, you go first.

Rob Strechay

>> Either way. But I think Paul and I are probably going to be very much in agreement on this anyways, so I think when you start to look at how this is transformed... I think it is transformative. I think what GenAI has done is opened up the ability to go to Agentic AI, and I think it really allows you to talk to your data and talk to your applications in a easier way. But I don't think it's the only one alone. I think you're going to still have all of what I put into the category of traditional AI that brings along and adds value to it because it's not simply about telling me how to cook a recipe or what ingredients do I need to do that or what have you, or travel even more complicated of using an LLM for travel planning. I think when you start to look at that, it's, "Hey, I want this to help me do travel planning. Not only tell me where I should go and prioritize based on top tens from all different data sources, but then help me go book it." And that's where you start to tie in the agentic technology and the chaining together. And I think, in that example, you start to look at it, this is really transformative to actually helping people be more efficient in their work or in their relationship with a company.

Scott Hebner

>> Yeah. Paul?

Paul Nashawaty

>> Yeah, Scott, I couldn't agree more with what Rob was just talking about, but one of the things I like to do is take it back a little bit. And one of the reasons I want to take it back is I think with regards to the AI, you mentioned hype, right? And I don't want to say that this is hype, but also want to transfer or move that word to maturity of organizations. Maturity of organizations will enter at different points. There are people in the infancy and there are people that are very, very mature. And what we find in our research is AI is an area that people are honestly scared about. When you're looking in the app dev world or DevOps world, the practitioners are going, "Oh, the AI is going to take my job." And it couldn't be farther from the truth. What we're seeing is AI is actually an operation efficiency tool, and what we're also seeing is whether you're using GenAI natural language or Agentic AI in the marketplace, we're seeing the adoption coming in at a number of different levels. And Scott, this is where I think it's interesting. When I talk to CIOs, their number one goal for CIOs or initiative I should say for CIOs is application modernization. And as we know, modernization is a treadmill. Once you're on it, you can't get off. But the two barriers that they run into is complexity and skill gap issues. In fact, what we're seeing... Excuse me. In fact, what we're seeing is in our research, 67% of organizations are hiring generalists over specialists, and that's largely due to the fact that they can't find the specialists. So AI and the level of maturity will help with things like automation, will help with transform the delivery and the market and also help them with these initiatives. But I also believe, Scott, that the professional developer will be focused on innovation more so than focus on maintenance with AI and AI tools and automation in place. So that's where I think it's going.

Scott Hebner

>> Yeah, I think you guys are right. The way I've pieced this all together is this is one big transformation that has been occurring for a couple decades. We first interconnected everything, then we instrumented everything. Now we're infusing intelligence into everything. And Rob, as you pointed out, we start off with the predictive models that can help us make predictions based on what has happened in the past. Then generative AI, I think creates that gateway into the world of AI, right? It's the equivalent of the browsers back in the early internet age democratizing everyone to be able to get to and use AI. And now going forward, it's how do you start to build on top and around the generative AI to create even more value? And I think what Agentic AI represents is more human and AI collaboration, and you get into achieving goals and helping to make decisions and solve problems. It really takes it from doing tasks into the world of, I need to achieve a goal, help me do it. And so I think it's this progressive and probably one of the most important transformations within AI. I think will be Agentic AI. Then of course, I'm sure in a couple of years we'll be talking about something else. But anyhow, one of the questions I've been asking everyone I've talked to is a simple one, really, right? What are the success factors in creating a winning strategy for this, right? I mean, there's so much going on. And I've summarized it into four key tenets that just from all the conversations. We'll bring up the graphic that show this. I think it starts with, first of all, the game is played in the field, not in the dugout. In other words, you got to start experimenting. You got to start doing things now. You can't wait because if you wait and you lag, it's going to be harder and harder to catch up. I think step number one is you got to adopt these new technologies quickly because the innovation cycles are just going at such a rapid rate that you really do risk falling behind. So don't wait to see what happens with the Agentic AI get involved. I also think it's important when you start to deal with the notion of coworkers and organizational workflows and outcomes and decisions and things of that nature that you codify the ROI. You got to be really crisp on what you're trying to achieve, and the best way to do that is to measure it. So that has come out loud and clear to me. The second thing is trust, right? It's all about trust. It's one thing to ask GenAI to help write something or to generate some analysis of information or automate a simple task. It's another whole thing to actually rely on it to make decisions or to help you make decisions and even autonomously go off and act. So trust becomes the currency of innovation. No trust, no ROI. And then, finally, keep in mind that this is all about empowering your talent, right? It's coworkers, not replacing them, but augmenting what they're able to do. So you have to have a talent strategy, which brings me to the fifth one, which is about culture. Culture is not part of the game. It's the entire game. As Lou Gerstner, the ex-CEO of IBM once said, culture is so important here because you're trying to transform how work gets done. And you can't do this in a silo from a technology perspective. So that's how I sorted out all the input I've gotten. And what do you guys think? Tell me a little bit about... You start off, Paul, this time.

Paul Nashawaty

>> Sure. Yeah, no, I think it's great. I think you laid out a nice awareness of how AI can impact the organization. One of the things that I hear quite frequently is there is an element of utilizing AI, but also, there has to be a level of dependency that needs to be understood. And what I mean by that is when organizations put out a code or put out applications or documents or whatever it may be with AI, one of the things I heard from one organization was, "Well, it's not our fault that the code was wrong. AI wrote it." And it's like, well, no, no, you actually have accountability. So I like, Scott, what you were talking about with the culture and embedding it. It has to be a part of your culture and it has to be treated right. I mean, it's like AI, it's a screwdriver and drill analogy. You have a hand screwdriver. You can put the screw in the wall, or you can use a drill, you get operational efficiencies, but if you use the drill to powerful, it'll put the screw through the wall. And that's the same thing here. You don't want to put the screw through the wall when you're delivering code or anything else. So compliance, regulations and governance is also key because one of the things that we're also seeing is with the proliferation of applications, the citizen developer is taking on more of a role in development, and AI is being used more and more at the lines of businesses, therefore the citizen developer needs to have those governance and controls in place as well. So yeah, Scott, I think that's really where this is going. It aligns nicely to your ROI trust and talent bullets.

Scott Hebner

>> Rob, what would you add or subtract or modify or what are your thoughts from what you found?

Rob Strechay

>> Yeah, I mean, I think you both did an awesome job there, and I love how you have trust at the center of everything because to me, it is about trust and trust in the models, trust in the data, trust in those folks around you, trust in the developers, is beyond trust in the tools. And I think without trust, then none of this happens. And I think that leads to some of the others that you have there being, again, and I like how you have guardrails on either side of it with the whole playing on the field and culture is the game because to me, it's a whole company. Companies are going through this. I talk regularly with companies and in fact, some of our other folks like Rebecca Knight who studies the future of work, we have these conversations about how it's changing the culture of those organizations. And to Paul's point, how it's changing the culture of development, how it's changing the culture of data analysis and data engineering. It's touching everything. And it's not just that the AI is doing that, but it's the building of AI and agentic systems that is ripping at the fabric of culture of some companies. And I think if you put the people first and the talent first, and you really focus on that, you can have better outcomes and you get to the ROI because I think you do smarter things, which allow you to get the ROI of AI, which has been tough for many organizations. Just throwing up a chatbot really didn't get them the ROI they expected to. So I think this is why this summit in general is so important because I think Agentic takes that LLM to the next level with the other AI and other different types of applications that are plugged in.

Paul Nashawaty

>> And Rob and Scott, I do want to add just one piece, Rob, to your comment, it's spot on. I mean, what I'm hearing from organizations is SLOs, service level objectives, are really the north star. And people were using AI to get to that north star to align to the business KPIs. So yeah, I just felt that that was a very important point to add in here because historically what we saw is very siloed and monolithic approaches within organizations, and now it's more coming together, and it has to, right? But let's not forget, and Rob, you've touched on this very clearly, there's a human in the loop. A human has to still be in the loop. And even though you have the big green button, somebody has to push it. And I don't think we're ready to let AI take over the world here.

Scott Hebner

>> Yeah, yeah. No, I think this is a big deal here. And Rob, Paul, as you know, we're going to be doing some primary research around the trust index when it comes to all this. And I think in the bigger picture, what's really happening here is people are moving from lower risk use cases to more business critical use cases that involve decisions. And achieving trust, it just needs to become the overriding success factor that you focus on. No trust, no nothing, right? I mean, every decision, every investment should be made and evaluated for how it contributes to trust and outcomes. While it starts with ensuring cross-functional involvement in your organization in the agentic strategy and enhancing the data quality and relevancy and all that, the bigger challenge to tackle would be the human factor. Business leaders, workers already lack trust in AI a little bit today, right? I think it's around half actually trust the outcomes that they're getting from their AI when they're making decisions. And with agents now promising to make business critical decisions and even act autonomously, the issue of trust would become even more pronounced. So if you think about it, our business is really going to just blindly trust AI agents? I seriously doubt it, and that's why I've hardened my view that trust becomes the currency of innovation as we go forward here. The more trust, the more business will be willing to buy into innovation and Agentic AI. And in the end, building that trust is what you need to focus on. You need to be able to, one, transparently explain in the language of the user, the agent has to be able to explain what it's recommending it does. The user has to be understand why one decision may be better than other decisions with a good amount of detail. And then, thirdly, allow users to intervene and play out alternate scenarios. So to me, this is a do or die imperative. You got to get that trust. Agree? And that's what you're hearing, right, collectively here?

Rob Strechay

>> Yeah, I mean, I'll jump in. I think, again, trust is the currency, as you put it, that we're going to be dealing in, I think, in this next year. And I think that organizations need to know and trust that the agent is not going to go off and do something crazy, like give away free airline tickets like what happened to one of the airlines when somebody did a prompt attack on it and was able to get a crazy discount for no apparent reason. I think that was mainly an LLM use case, but you start to look at it and as the agentic gets bigger, and because it's more than just an LLM, you're going to have... It's going to be more complex. And that trust in all of the pieces and how they're wired together and what platform they lie on is going to be key to where people are going to spend their money. And I think where organizations are going to spend their money is where they can get that trust, they can get the guardrails, they can get the ability to solve for some of those cultural issues that will make it simpler for people to develop on the platform for add pieces to the platform to make it. Because certainly with the volatility that's been going on in the markets over the past month or so, I don't want my AI going out there that it hasn't seen somebody go and do a whole bunch of tariffs. That to me is when it starts to take into things and into account things that really it hasn't seen, how do you do that? Having the human in the loop at that point makes a lot of sense because otherwise there's still some things that AI can't foresee or has not seen that could change the dynamics, and actually we can trust with that end user customer of that investment house, for instance. So yeah, I think trust being at the center, as I said earlier, before even knowing you were going down that path there, I think, again, you hit it on the head.

Scott Hebner

>> Yeah, you think about it, this is going to require a ton more sophistication than what generative AI and LLMs today can deliver on their own. I think they're foundational, right? This is especially true when you get these AI agents to work together to collaborate within agentic systems and workflows where you got partnerships, building within some kind of work between a human and a digital worker. And I think the good news here is more and more vendors, more and more companies are democratizing all these capabilities. What I've pulled together here and with the help of all you and the people we've been working with in terms of companies, we've come up with this metaphoric ladder to Agentic AI. And George Gilbert was a big piece of this, which provides a simple headlight into the journey ahead. And let me bring this up and let's just talk a little bit on the technology roadmap because I think you're going to hear about each one of these during the sessions that make up this summit. That's why I encourage you to watch them all. Generative AI, the foundation, and the LLMs, the predictive models, that's what's going to give you the gateway into the intelligence of AI. Step one is going to be you have to deploy domain-specific knowledge models that really understand the domain that you're trying to make decisions within. The context of the entities and the data set, how they relate to each other, very, very important. The context, the perspective that has to be built into this extended set of models on top of an LLM. You then have to infuse explainable decision intelligence, right? The decision intelligence has to be built in because it does not come from generative AI. And again, what good is a decision if you can't explain it, because if no one can understand why, they're not going to trust it. Then from there, after you build that extended ecosystem of models, then you got to build your agents. And our research shows more than half of businesses plan to actually acquire pre-built agents, customize them, and then integrate them with their own proprietary agents that they build. So you're going to need a platform to do that. And then, of course, you get to the top layer here is how do you start to wire these agents into a system that can accomplish more complex work or help more complex decisions that a single agent can't do on its own or when it's working with a bunch of people within some process to accomplish an organizational goal? And then this whole system has to learn continuously. Each one of these you see on this so-called ladder to Agentic AI, we're going to touch on in this summit by different vendors and some of the customers that we've heard. That's the good reason why you want to watch every single session, right? So Rob, Paul, any key points you want to emphasize? Paul, let's go with you first here.

Paul Nashawaty

>> All right, yeah, sure. I think that, one, I love the maturity model that you have laid out here. I do want to say I like you used the word ladder, but I also think that any of the organizations looking at this can enter at any stage of the maturity. So it's not just the ladder. They can be on a couple of rings up and they can get in there. The importance of this summit is incredibly important because of a couple of factors. And Scott, I know you've been echoing saying, "Hey, you want to watch every episode and every part of it. And I agree. And here's why. Because, one, AI in organization, as I mentioned earlier, has a separate budget in many organizations. Two, it's a competitive advantage for a lot of organizations. And if they're not using it properly, they can be competitively boxed out of the market. But three, this is where I think it's incredibly important to think about for your organization, for the viewers watching today. It's incredibly important for you to think about not overspending or overpivoting on AI initiatives without fully educating yourself on how are you going to use it. And that's where this summit will help educate you on what your journey looks like and that model, that ladder that you showed, Scott, gives the steps on how to get there. But again, understand from the audience perspective where you are in your own journey in order to get there. So I think it's incredibly powerful and I think it's really well-timed because I think budget right now is people are ready to spend, they just don't know what to spend on.

Scott Hebner

>> And it goes to that whole notion of the game's played on the field, not in the dugout. Start playing around, start researching. Rob, what are your thoughts?

Rob Strechay

>> Yeah, no, I love how the ladder is. The first rung is understanding the knowledge models and looking at that because if you don't know your processes, and I think process automation has been around for a long time. It's gotten very sophisticated over the last few years. I think it's that first rung on that ladder. I know I've talked to George many a time about it. We sit there and you start to look up from the process and then say, "Okay, what are the steps that I can automate that AI can help with? Where does the human come in? Where do I not have to have a human?" and understanding and mapping that out so you can get people out of the repetitive stuff that really kills productivity. And I think to know that and to understand that it's not all about the tech, but the tech enables you to go faster in doing these things. And having a platform to go and do and build agentic technology on top of really will help you absolutely leapfrog a couple of rungs potentially to Paul's point and help in your maturity model as well. So I'm really excited about all the speakers that are coming on today because I think they all have really unique perspectives on how to go about this. And I think that's the key is don't just listen to one voice. You got to listen to a number of voices and understand how it fits your journey to agentic.

Scott Hebner

>> Yeah, no, those are some good points. And just think back to the last three years, right? Gen of AI came on the scene three years later. We're very much in the infancy of Agentic AI and it's going to mature rapidly, but I think we're still in that early stage. So it's time to experiment. We don't know what's the absolute right approach to getting this to work and to get to ROI, but it's time to get involved. All right, Rob Paul, really appreciate you being here. This has been an awesome conversation. All of us will be giving our analysis and keeping you guys up to date on social media and LinkedIn and so on. So please watch for all that. Please make sure that you all start to watch the sessions in the AI Agent Builder Summit. Again, there's 12 of them, and you'll hear directly from those leading the charge in Agentic AI. And we bring up a chart just to show you who's all you're going to hear from here. You'll gain valuable insights and practical advice from these companies, from real world implementations and real world business strategies. You won't want to miss, as we've been saying, any of these sessions. You can access the agenda in all these sessions by visiting the summit portal on thecube.net or by checking out our YouTube channel. And on the summit portal, we can also learn more about the participating companies and the leaders speaking across these sessions. Thank you very, very much for tuning in. We appreciate it. We'll see you on the next session. We are the leader in enterprise tech news and analysis. Bye for now.