Exploring Innovations and Key Insights from the AI Agent Builder Summit
Jayeeta Putatunda of Fitch Group, director of the AI Center of Excellence, Paul Sciaudone of HP, vice-president of platform engineering and quality, and Tim Sanders of G2, vice-president of customer insights and executive fellow at Harvard, participate in the AI Agent Builder Summit at theCUBE.net. They explore the evolving landscape of AI agents, discussing industry-specific applications and challenges.
In this engaging session, Putatunda provides insights into the financial services industry's nuanced requirements, emphasizing the transformative potential of AI in managing complex, unstructured data. Sciaudone highlights advancements at HP in software development cycles, explaining how AI agents enhance productivity and innovation. Sanders offers a broader view of the marketplace, emphasizing the widespread impact and applications of AI agents across multiple sectors. Hosted by Scott Hebner from theCUBE Research, the panelists explore the practical realities and opportunities AI agents offer.
Key takeaways include the importance of trust, governance and collaboration between humans and AI agents for optimal results. The participants stress the need for standardization to ensure security and reliability in AI deployments. According to Sanders, leveraging AI must be a strategic move aligned with business pain points, maximizing technology's potential while fostering organizational trust and skill development.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
AI Agent Builder Summit, Encore Presentation. If you don’t think you received an email check your
spam folder.
Sign in to AI Agent Builder Summit, Encore Presentation.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open this link to automatically sign into the site.
Register For AI Agent Builder Summit, Encore Presentation
Please fill out the information below. You will recieve an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for AI Agent Builder Summit, Encore Presentation.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
AI Agent Builder Summit, Encore Presentation. If you don’t think you received an email check your
spam folder.
Sign in to AI Agent Builder Summit, Encore Presentation.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open this link to automatically sign into the site.
Sign in to gain access to AI Agent Builder Summit, Encore Presentation
Please sign in with LinkedIn to continue to AI Agent Builder Summit, Encore Presentation. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
AnalystANGLE Customer Panel
Exploring Innovations and Key Insights from the AI Agent Builder Summit
Jayeeta Putatunda of Fitch Group, director of the AI Center of Excellence, Paul Sciaudone of HP, vice-president of platform engineering and quality, and Tim Sanders of G2, vice-president of customer insights and executive fellow at Harvard, participate in the AI Agent Builder Summit at theCUBE.net. They explore the evolving landscape of AI agents, discussing industry-specific applications and challenges.
In this engaging session, Putatunda provides insights into the financial services industry's nuanced requirements, emphasizing the transformative potential of AI in managing complex, unstructured data. Sciaudone highlights advancements at HP in software development cycles, explaining how AI agents enhance productivity and innovation. Sanders offers a broader view of the marketplace, emphasizing the widespread impact and applications of AI agents across multiple sectors. Hosted by Scott Hebner from theCUBE Research, the panelists explore the practical realities and opportunities AI agents offer.
Key takeaways include the importance of trust, governance and collaboration between humans and AI agents for optimal results. The participants stress the need for standardization to ensure security and reliability in AI deployments. According to Sanders, leveraging AI must be a strategic move aligned with business pain points, maximizing technology's potential while fostering organizational trust and skill development.
Vice President of Platform Engineering & QualityHP
Tim Sanders
VP Research InsightsG2
Jayeeta Putatunda, lead data scientist and director for AI center of excellence at Fitch Group Inc.; Paul Sciaudone, vice president of print cloud platform engineering and quality at HP Development Company LP; and Tim Sanders, vice president for research insights at G2.com Inc., join theCUBE’s Scott Hebner at the AI Agent Builder Summit for a customer panel on real-world applications of agentic AI.
The discussion highlights sector-specific perspectives on AI deployment. Putatunda explores the complexities of applying AI in financial services, especia...Read more
exploreKeep Exploring
What are the backgrounds and perspectives of the panel members discussing agentic AI at the summit?add
What are some benefits and challenges of using generative AI in handling complex unstructured data in financial services?add
What advancements are professionals in the field of AI hoping to see in the coming years to make their jobs easier and to create more powerful AI agents?add
What is the importance of standardization and interoperability between agents and platforms in the development of agentic capabilities for enterprises?add
What is the importance of trust in businesses and end users when it comes to agents helping people plan, make decisions, and solve problems?add
What is the importance of understanding the underlying process and workflow in decision-making, particularly in the financial sector, in order to build trust and come up with innovative solutions?add
What approach should organizations take when implementing artificial intelligence technology?add
>> Hello everyone. Thank you for tuning into the AI Agent Builder Summit here at thecube.net. This is where we're going to be exploring best practices and proven solutions that can help you build AI agents and agentic workflows, and, more importantly, achieve trust in outcomes and accelerate your ROI. I. Am Scott Hebner, the principal analyst for AI at SiliconANGLE Media in the Cube Research, and I'll be your host for the summit. During the summit, we'll hear from industry analysts and a cross-section of industry-leading solution providers such as Semafor.a, AgilePoint, Ascendion, Deloitte, Geminos, IBM, and SAS Institute. But before we hear from them, we felt it was important to understand the perspective of end user organizations and the overall marketplace, their aspirations, their needs, their challenges, and their approaches to agentic AI. So I'm very excited today to have with me a great panel for us to have this discussion. First, Jayeeta Putatunda, she is the director of AI Center of Excellence at Fitch Group, that's on Wall Street, who has a very strong point of view on the evolving needs of the financial services industry. And next we have Paul Sciaudone, the vice-president of platform engineering and quality at HP. He is looking to evolve their SDLC, leveraging agentic AI development capabilities, common tools, practices across a complete end-to-end development life cycle. And then Tim Sanders, who is the vice-president of customer insights at G2, the world's leading software marketplace, and an executive fellow at Harvard working with their Digital Data Design Institute. He is a top-of-class student of the AI marketplace. All right, well, I'm excited to get this discussion going. I suspect we're going to hear some brutally honest assessments of what's really happening on the ground in real businesses, and what it's like to actually be accountable for the outcomes as they start to invest in build out AI agents and agentic AI initiatives. So let's dive in. Paul, Jayeeta, Tim, thank you very much for being here.
Paul Sciaudone
>> Great to be here.
Tim Sanders
>> Great to be here.
Jayeeta Putatunda
>> Thank you.
Scott Hebner
>> All right, thank you. We're going to have some fun today. So let's start with what excites you about the potential of AI agents? That is, why are you pursuing this strategy? And let's start, Jayeeta, with you. You're representing the financial services sector. What's making you excited about this movement?
Jayeeta Putatunda
>> Absolutely. So much to be excited about, right? Think about how generative AI pushed us to do so much that we initially thought was not possible, especially with complex unstructured data, that we handle in financial services. If you look at a PDF, all PDFs look different, variations of PDF in the kinds of document we use, like 10-K annual report, ESG, everything looks different. So generative AI really helped us to kind of do the summarization, the final layer of generation. But what now the agents are helping us do is basically help with the retrieval, because we all know retrieval was tough, connecting all the knowledge bases, web searches, different variations of data points that we might have. I feel this is like, we are just making sure that the kind of data, or again, making sure that there's no garbage in garbage out problem that we are still seeing with generative AI. We have to have the incorporated data into the system really, really clean, and really nicely retrieved. And that's what I feel some of the tasks that I've been working on really helps in getting that data in the right way, in a more automatic way, as well as making sure that we keep the quality really lean and clean. So very excited about, we can keep talking more about the use cases, but I feel I would love to hear what some of our other speakers have to think about it.
Scott Hebner
>> All right, well Paul, what excites you? Let's go to you next.
Paul Sciaudone
>> Yeah, I mean, what's really exciting is agents are really taking software development to the next level. How you think about building applications, and services, and all these capabilities where it really used to require core software development and coding and development to do that, where you've now got an ecosystem of agents that essentially become the Lego blocks to build new capabilities and new services. And you think about prototyping, getting software out there quickly, you don't even need to be a software developer to do it with agents. You can get product people building things quickly, rapid prototyping, even our customers can work with us to prototype together. So it's really taking what we used to think about conventional software development to the next level. So it's really exciting. It is almost up-skilling everything across the industry.
Scott Hebner
>> Yeah, it superpowers. I recently put out a research paper and we're projecting that software developers that are going to become data scientists will grow 30% compounded over the next five years. And it's because they can start to adopt these agents, and the agents will do a lot of the data science and AI engineering work, and it's almost like it fuses into a joint profession. You're a AI engineer and a software developer, and it's because of the agents, right? Because there's not enough human talent to keep up with all the data that's out there. Tim, what about you? I know you're super excited.
Tim Sanders
>> I'm excited about it, yeah. Somebody said, "What do you cover? I said agents, agents and agents." Agents solve the delivery problem that generative AI doesn't. Here's what I mean by this. Generative AI is a suggestion machine, and all of the research like BCG, Harvard Research suggests you can be 30% more productive, but that doesn't mean that the productivity is delivered to the organization with more velocity. Because there's a great concept called Parkinson's law, and that is people will always stretch their work schedules to the time that's given to them. So even if a coder saves 30% with generative AI, that doesn't mean they produce 30% more content, but with agents, they do. Agents don't take a day off, they don't get distracted by TikTok, they don't go on PTO, they work 24/7 to deliver. They're not suggestion machines, they're action machines. And what I'm seeing now is, whatever your use case, customer service, sales, software engineering, you can literally change quotas and forecast reducing debt with agents. You really couldn't do that with generative AI. It was more Waze, agents is more Waymo.
Scott Hebner
>> An interesting perspective, Paul and Jayeeta. That does make sense to me, just because you can do things 30% quicker, doesn't mean humans are going to do it 30% quicker, because then they just delay. Yeah.
Jayeeta Putatunda
>> I really like that point, what you highlighted, Tim, but the concept is how do we make sure that we are also doing it? I think it's a problem with the financial industry. We have a lot of guardrails and governance that we have to maintain and go through. So we also have to make sure that we're not completely optimizing the agents. It's absolutely agentic AI, but with a lot of layers of guardrails, human in the loop review, as well as making sure that we are always grounding it in some level of ground roots. Even if it's hard together, sometimes there's no other way to do it.
So I think it's a great exciting world, but we have to kind of balance it just to make sure we sell to our clients as well as leadership, a trustworthy solution, not really with a little bit of transparency.
Tim Sanders
>> I would say one thing, I talked to a lot of people in financial services as well as healthcare where this also comes up, right?
Jayeeta Putatunda
>> Yes.
Tim Sanders
>> And I believe that the challenge is also the opportunity. I feel like Jeff Bezos saying, "Your margin is my opportunity," but what I mean is historically in digital transformation, financial services and healthcare has always kind of lagged a little bit, because they have to be so darn careful with the data. And that's why I see such disruptive opportunities. I think of Ant Financial in China, and how they absolutely disrupted financial services a decade ago with automated loan approvals in a world where the Wells Fargos and BOAs are like, "We'll never do it." So I think it's an interesting challenge as well as an opportunity.
Jayeeta Putatunda
>> Yeah, absolutely. Definitely.
Scott Hebner
>> All right, well let's kind of take a step forward here. I'm just curious to the use cases in your business that are now possible that weren't really possible before compared to, for example, the discussion around generative AI. And you touched on it a little bit, but could you be a little bit more specific about what you think could be done now that just simply couldn't be done before? And why don't we start, Paul, with you on that one?
Paul Sciaudone
>> Of course, even at HP, we've been using some of the generative AI capabilities of GitHub, Copilot, Cursor, you name it. And those definitely help developers become more productive. What we're seeing now, what's exciting is when you look at the complete SDLC around having a product manager describe a business case or a use case, and some of the capabilities of some of these agents and these tools can actually take those requirements and break them down into story points and epics. They can start creating the code for you, which that was never even available 12 to 18 months ago. So for us, it's really looking at, how do you take the complete SDLC as you have it today, and how do you optimize that? And in such a competitive market around new product development, new feature capabilities, and some of these companies taking 18 to 24 months to launch a new software product out there, you can reduce that time dramatically, and by at least 50%.
And that isn't just writing code and pushing it out there, it's doing the business case analysis. It's writing the user stories, it's doing the coding, it's doing the testing, it's doing the development. So we're really looking at, how do we apply agentic AI across the SDLC to really move the industry fast-forward, and come up with new solutions that we're never capable of before. So for me, that's what's exciting, is we're looking at how can we be disruptive, not even just internally, but in the market and getting new solutions out there quickly.
Scott Hebner
>> What about you, Tim?
Tim Sanders
>> First of all, I just want to plus on what Paul said, that this idea of getting software written ahead of schedule, on schedule is really important. When you talk to a lot of people in enterprise, the common story is that software releases are typically late to market, sometimes three, sometimes six, sometimes nine months, because pf the human factor that agentic seems to address. If you look at this chart, researchers say that if you're six months late to market, you lose 30% of the lifetime profits of a software product, because the market doesn't wait for you. You miss the peak activity in that market. So Paul makes a really compelling point for what I think is one of the more important use cases right now in agentic, that CFOs can get excited about. Now at G2, we have an agent now, Monty, that previously launched as a chatbot, but now Monty has access to tools, whether it's information retrieval, scheduling, et cetera. And what Monty's able to do is create a dynamic automation solution for software buyers on G2, that can literally take them from query to canvas, introduce them directly to the vendor to scale, and we're excited about that. Many of the customers I talked to are identifying three use cases. Paul mentioned software coding. That's one. The most common use case I'm hearing now is customer care, especially in those organizations where customer debt is really, really high. But an emerging one we're seeing right now in the research is around sales coaching and sales development. We're seeing pitches in the Silicon Valley for companies that are projecting $200 million ARR with 10 employees and an army of agents. And that's a few years away, but incredibly exciting.
Scott Hebner
>> Wow, that's incredible. And Jayeeta, I mean, obviously in financial services you got rapid cycles, you got a very dynamic set of conditions and marketplaces around the globe. Getting software out is really important, right?
Jayeeta Putatunda
>> Software out, as well as helping our internal analysts make sure that they are able to shift to thousands of pages of complex PDF really, really fast. So how do we break down compliance reporting, take as an example. How do you make sure that any new rules that's coming out quickly, real time, evaluate and see what stages we are in, what the gap is, and then suggest very different unique solutions that we can go and address it. I think another really cool use case that's really, really helpful that we have been seeing in the financial services is very quickly breaking down huge documents into multimodal features. Because sometimes I know we have multimodal models that really help us break it down, but when you're talking about, I don't know, batches of thousands going in for a particular topic, it really gets hard to process it in real time. So how do we make sure that the back end document review pipeline is really categorized? So agents are helping us break down tests, images, infographics, then go each along the data pipeline route and then evaluate on the go, and make sure that everything is uploaded or shared in the knowledge base, for our analysts and the back end users to really leverage the data really, really fast as soon as some new PDF is published in any of the things that we monitor. So a lot of amazing use cases, and I'm really excited to see how we can build it again in a responsible way, making sure we have enough evaluations, enough guardrails in place, and adding to the governance layer that we all should uphold for any of the solutions we build.
Scott Hebner
>> So as you guys look over the landscape of technologies and solutions that are available out there today, what advancements are you hoping to see in the coming years to make your jobs easier, and to create more powerful AI agents? I mean, there must be some gaps that you see today, and we're just early in this whole agentic journey. So what do you want to see coming next? Tim, you endlessly study customers, behaviors, needs, challenges. What are the couple areas that you think are going to emerge that are really needed out there?
Tim Sanders
>> Well, first of all, we just have to address the reliability of the underlying language models, right? Because when you think about agentic guardrails, the guardrails get wider as the reliability of the underlying technology gets better. And I'm incredibly excited about model context protocol. I think that's really important in terms of developments allowing the agent to go find the data instead of depending on retrieval and traditional RAG, because human in the loop is a short-term solution that creates a long-term problem for organizations in this agentic economy. So I'm excited about this idea of MCP being the new unlock for enterprises over the next few years. Stay tuned for that one. The other thing I think that's going to be important, and I was just reading fascinating work from Google on this yesterday, is the concept of A2A, that's agentic to agentic communications. In the future, agents will talk to agents and the current communication signals we have in place are going to break for most organizations, because outside agents will need be contacting us to scale, especially in customer situations. So I think as we think through more of A2A communications, we're going to see a lot more scale in the future. And then to the point of what's been brought up earlier, we've absolutely got to do a better job of considering GRC and security up front, because agents only work to scale if they have access to external tools, and that's going to require, at least our research says, some serious legwork on both security as well as GRC.
Scott Hebner
>> Interesting. Jayeeta? ?
Jayeeta Putatunda
>> Yeah, I mean, I completely agree with Tim, and I think the biggest area ... And I have always felt about this, even for generative AI, that sometimes it gets really black boxy. How do you make sure that even if you're doing red teaming and making sure that all the ... You're not passing your data to any model that you don't have really insights on what it was trained on. Similar for agents, when we break down agents and agents are communicating with agents, like Tim said, with A2A and MCP, it's going to get a little more autonomous. So how in the financial services, how do we handle this? How do we make sure that our data is safe? Our end customers are not really getting anything that would hurt us in terms of reputation, as well as cost. So how do we make sure that there are some intermediate steps, even if it's human in the loop for short term, or some kind of evaluation checks. So I think evaluation companies or startups who are really building these solutions, or even big tech, who's coming up with some of these frameworks, really need to think through for all kinds of data sets. And we have data sets that are really, really private. So how do we go about and make sure that the authenticity, the privacy is maintained with some of these technologies.
Scott Hebner
>> What about you, Paul?
Paul Sciaudone
>> Yeah, and just building upon what Jayeeta and Tim said, right, I think it's really going to come down to more standardization and interoperability between agents and platforms. Like Tim said with A2A just being announced earlier this week by Google, I think that's a huge leap forward as companies look to integrate a lot of these agentic capabilities and make it easier to integrate. I think what's been happening in the past, I think, 12 to 18 months is everyone's been rushing to get out there first, and build out these new agent capabilities, and now with that proliferation needs to be some sort of standardization to build that trust and security, and make sure it is reliable and standardized, so enterprises can actually rely on building agentic capabilities, that isn't going to change three months from now or even sooner. It's building upon the future. I also think it's important, as we look at making it easier to integrate with humans in the loop, I think security and trust is clearly front in mind, but I think the more you can integrate humans in the loop and provide oversight in those tightened guardrails is going to be super important. Is there ever going to be a world where it's completely agentic? Possibly, right? Was it going to require humans in the loop? Absolutely. And I think depending upon ... Jayeeta, I see you shaking your head. Financial industry? I completely agree. It may never be completely agentic. But on the web application and consumer side, possibly. I think it really depends upon the risk profile of each one of these verticals, and what customers and people are comfortable with, right?
Scott Hebner
>> That's-
Paul Sciaudone
>> So Jayeeta, I agree. The financial-
Jayeeta Putatunda
>> No, I, absolutely. I was just answering from my perspective and I completely-
Paul Sciaudone
>> Yeah, it'll never be. But again, I think it's going to depend upon the industry, and the market, and the adoption rate of some of these, whether it's an end-user customer or the enterprise.
Scott Hebner
>> Well, let's play a game show, just yell it out, whoever's first. What do you think my answer would be about an advancement that's needed? And this is unfair to Paul-
Jayeeta Putatunda
>> Causality in agentic AI.
Scott Hebner
>> Yeah, yeah. No, I'm a big fan of the need for AI decision intelligence, like real decision intelligence, because the LLMs and generative AI and a lot of these agents, a prediction is not a judgment. To make a judgment, you have to understand consequences. Consequences are a function of understanding cause and effect, and a decision is a series of judgments that you need to make. And there's emerging technologies like causal knowledge graphs, causal AI in general, knowledge graphs in general that are a little bit more dynamic. I think there are some really key technologies that are getting democratized today by a variety of vendors, and we'll be having one of them here at the summit, AI that knows why and can help you not understand what to do, but why to do it, and how to do it and why certain decisions are better than others. And it kind of gets to the explainability issue. So that's what I think needs to develop in the years ahead. Otherwise, these agents are just not going to be able to make decisions reliably.
Jayeeta Putatunda
>> More reasoning through historical as well as more contextual data, like you said.
Scott Hebner
>> Yep. And I know Jayeeta did, where was it, the AI summit where you did a session on CausalRAG, right?
Jayeeta Putatunda
>> Yeah.
Scott Hebner
>> As an example.
Jayeeta Putatunda
>> Absolutely. Definitely adding ... Use it more as a tool to make your rag a little bit more powerful.
Scott Hebner
>> That kind of brings us to the trust factor, right? It's one of the themes of our summit. Trust, I believe, is the currency of innovation going forward. No trust, no ROI. How important is the trust factor in your businesses and your end users? I mean, if agents are going to help people plan, and make decisions, and solve problems, what's going to allow the users, the humans, or the co-workers feel confident on what the agent's actually recommending? Jayeeta?
Jayeeta Putatunda
>> Yeah, like I said, understanding the underlying process and workflow that the agent is using to break down the tasks, having a very clear understanding of how it went to that decision, so like, say, reasoning. The way that we feel, or at least I feel very strongly in the financial sector, is that we have to make sure that we are ... Again, trust is only then built when we make the human part of the loop really comfortable with the solution. And how do we make it? By upskilling them, making sure that they understand the technology, as well as what the limitations are. The moment they understand the limitations, it'll help them think through the solution much more creatively and out of the box. And then, that would help us to get to the right solution faster with the help of the agents as well as more creatively with the help of the humans. So honestly, if you ask me, I would not wish for a future where there's only agents and no humans, but I definitely feel that we can automate a lot of the manual tasks so that we can do more creative thinking, more creative development, and make sure that the human in the loop or the folks actually understand the technology loopholes, and help us build better solutions and work flows.
Scott Hebner
>> I would imagine, Paul, that's really important in the software development life cycle, because software development is sort of a combination of an art and a science. So the humans, what do you think?
Paul Sciaudone
>> Yeah, I mean, I agree, and I think even to my point earlier, I think the more we see around standardization in protocols, and a lot of these capabilities, and the focus on security, I think that's really what's going to help kind of build some trust. And I think a lot of it is just, it's still so new, and it's advancing so fast, it's really hard to pinpoint what is it going to allow that trust. But I think as you see some of the larger agent and agentic and generative AI players, the more that they standardize in the industry, and others will follow, that will definitely build some trust. I also think the more this gets adopted by companies and the users of it, it becomes part of their natural workflow, and that human in the loop, I think it's just going to take time, right?
Tim Sanders
>> I think you're right.
Scott Hebner
>> Yeah. I think that's, I kind of equate a lot of this. You're around long enough, you start to see the same movie over and over again with just different characters, and it was sort of like the early days of the internet age, where all the browsers had different protocols, and the value, turning into a gateway into the internet is when the browser is actually standardized, and hopefully the gen AI stuff at least begins to standardize, where then you guys can build on top and around those AI gateway services, if you will, and that's where the business value is going to come, I think. Tim, what are your thoughts on that and the trust factor?
Tim Sanders
>> As my sparring partner, you know I'm thinking a lot about trust right now. You have to ask yourself, how important was trust to cloud adoption? Well, $44 billion of deadweight loss over 15 years tells a story, that if you don't trust cloud, you don't migrate, or you wait. So I think it's fascinating in the world of agents, there's a great book for everybody to read. It's called the Technology Fallacy. It says, when you study why companies don't become digitally mature, it's never the technology, it's always the people. So when I think about trust, I think there's two types of trust. There's mental trust that comes from improving the technology, making it more explainable, all these other things we've talked about. And then there's physical trust, and you see that with Waymo. There's research now that says you take a person who says, "I don't believe in self-driving cars," and you just get them at a convention to get into Waymo in San Francisco and go a hundred feet, you've made incredible progress on their trust factor of Waymo. So my prescription to organizations is that every employee should be spending four hours a week, hands on keyboard, using current generative AI, or building their first agent and deploying it, so they get that physical level of trust in agents. Because I believe in this world, that's going to be the last element that's going to unlock companies, especially legacy incumbents, who I think will be the most at risk.
Scott Hebner
>> Yeah. All right. Well look, last question before we have to start to wrap up here. What would be your advice to organizations that are just getting started with agentic AI? They're just starting to think about how to build agents or acquire agents, and customize them or whatever, but just they're at the very initial stage of their agentic AI journey. What would you advise them in terms of getting started, and then what are some of the key success factors that they should really be tracking? And Tim, let's start with you on that.
Tim Sanders
>> Start with the pain. My advice to organizations always when it comes to artificial intelligence, don't start with technology, start with the business and work backwards. What I'm seeing now is that the early success factors for agentic, and, for that matter, have been so for generative AI, has been utilization, but utilization usually follows pain. That's why in G2's recent survey, which we haven't released yet, it was just done a few weeks ago. When we look at organizations that forecast adoption of agentic to scale, it's always because they started with the use case where they had the greatest debt, which means they had a tolerance for some error and some growing pains. That's why you see customer care as being, percentage-wise, the most common use case right now for the agent builder platforms. So follow the pain and work backwards. You'd be surprised how much more forgiving we are of technology.
Scott Hebner
>> Jayeeta?
Jayeeta Putatunda
>> Yeah, absolutely, I agree to that. And I think of it as like a problem-first approach. I know there are so many frameworks that comes out every day and we we're like, "Oh, let's try this model, this framework," but think about what problem, like Tim was saying, you need to solve. And another area is communication. I feel sometimes people are so, like Tim was saying, why do they not physically accept or trust the system? It's because they don't know about it, or they don't know its capabilities. So I think the communication chain from the leadership and whoever, the technical people that are actually working, it's our kind of responsibility to bring them along the journey and talk to them about how it can help their scenarios, how it can ... Or how they can become more knowledgeable around this technology, share resources, help them upscale. I think once you do this in a more, I would say, do it together, you'll be able to, I feel, add more value and bring your entire company up at the same time. So yeah, problem-first approach as well as bringing people together to upscale with you. That would be my advice.
Scott Hebner
>> What about you, Paul?
Jayeeta Putatunda
>> Yeah, my advice is, look, it's here, right? Hop on board. There was a time, well, it still is, right? Software was the future. Software was eating every industry. It was the next revolution. Software is now eating software, and it's now we're at a step function of it's here, find a way to apply it, because if you don't, somebody's going to. So it is a race, and it's happening really fast. It's happening all around. Find a way that you can apply it within your business, in your market. Your customers are going to expect it. It's really going to be the key differentiator when you can compare yourself around your partners, your peers, other companies. There's amazing resources out there, I know between vendors and research. We're at a new stage now where this is all accessible. It's easy, right? The agents will actually train you, they'll do the work for you. So find a way that you can embrace it, but you have to embrace it. It is here.
Scott Hebner
>> Yeah. I'll build on that by just saying, actually, I had a conversation with Irving Wladawsky-Berger, he's known as the father of e-business at IBM a long time ago, and drove probably one of the biggest turnarounds along with Lou Gerstner there. And what he would always say is, "The game is played on the field, not in the dugout. Start putting some points up, and start slow and build." And I think that's right on. It's time to get involved. If you're not already. It's really time to get going. All right, Jayeeta, Tim, Paul, thank you very, very much. This has been an awesome conversation. I'm sure the audience has learned a lot. I know I certainly did. Thank you so much for being here, it's super appreciated. For all of you, please make sure you check out the rest of the sessions that make up the summit to hear from industry analysts and solution providers. You'll learn a great deal more about the best practices and the new technologies that can help you build and deploy AI agents, and how to differentiate them, by building agents that can reason alongside human coworkers, so that your organization can make better, more timely decisions. And with that, make sure you go out to the portal that is on thecube.net. You'll learn there about the other participants in our summit, and what other sessions are going on here. And that will be your resource to access the sessions that are going on, clips, articles, and all the different assets that come from everything that we're learning here. So please make sure you tune in, and we'll see you at the next session. Thank you very much. We are the leader in enterprise tech news and analysis. We'll see you again soon.
Scott Hebner
