ADI KURUGANTI

>> Craig Gould: Adi Kurungati, thank you so much for joining me today on the Master Move podcast. Adi, you are the Chief AI and Development Officer at Automation Anywhere, the number one cloud, native agentic process automation platform driving digital transformation for more than 5,000 enterprises worldwide. And you previously served as SVP and GM of Salesforce’s digital business platform. Adi, I want to talk to you all about Magentic AI, the work you guys are doing at Automation Anywhere, what you see, being deployed and what you see on the horizon. But I love to start these conversations with a common question, which is Adi, what are your memories of your first job?

>> Adi Kurungati: Correct. Nice to be here. thanks for inviting me. so what’s the. My first job out of college? Well actually I had a job during college in a chemical factory back in India.

>> Craig Gould: Oh wow.

>> Adi Kurungati: Looking at how soaps are manufactured, that’s typically not an experience when you’re in college. But I got to play with some chemicals but also be on the production line, as the soaps were getting packaged and moving them down the kind of the operating floor. So that was my first first job, during college.

>> Craig Gould: Anything you take with you today from, from that first job?

>> Adi Kurungati: I think the So for one it gave me a view like as consumers we look at the end product, namely soaps. But there’s a lot that goes in into making a great bar of soap. Just like there is a lot that goes into making and building an amazing product or a solution that customers adopt. And whether it’s the right type of chemicals that are put together, whether it is the, you know, the right packaging so you know, targeting the right, right customer base, but also the right pricing, that aligns with that. So there’s the entire process that goes into it. And it’s no different from how you build product, which is how do you think about the right buyer, how do you think about the outcomes that you’re going to help that buyer achieve? Then what goes into building a differentiated offering that really helps the buyer, achieve those outcomes? And definitely Agent tki, as we’ll talk about it, is a key kind of ingredient in making that happen.

>> Craig Gould: You’re currently, as we mentioned, chief AI and Development Officer at Automation Anywhere. And process automation is not something new in this market. It’s not something new for Automation Anywhere. But as the, the company has grown and developed, the utilization of AI into that process and into the product has, has grown and has really developed. Can you kind of talk about the history of the organization and how your offerings have transformed based on the technology that has evolved over time.

>> Adi Kurungati: Yes, I’m happy to do so. Obviously like you mentioned, process automation, frankly 40, 50 years, you always had some level of process automation even on the operating floor. So Automation anywhere is more than 20 year old company, went through various avatars as a company, as a startup, as startups usually do, really created this category around robotic process automation. I want to say around 2013, 2014, that’s when kind of the categories was created. And the entire premise behind robotic process automation was how do you use, how do you automate repetitive manual tasks so that you can enable humans like you and me to focus on what’s valuable. So but there’s a lot of focus on obviously cost reduction and fte, reduction during that phase. And it went through incredible growth. You know automation Anywhere, along with a couple of other vendors, ah, really benefited from the rapid acceleration let’s say in the 2010s. and to be clear, while we were primarily using deterministic task automation, there’s always a element of AI or AI ML, right? So for example even looking at a UI browser and figuring out the DOM structure and accurately predicting what needs to be automated, we had AI ML models there, including computer vision models, to look at it. So it’s not that AI ML is new to us. we’ve had AI, many of our products, even before the advent of generative AI and agentic AI. Then from RPA it expanded to intelligent automation which is really expanding the scope of use cases to processing documents, to mining process logs to figure out what to automate. There’s just an expansion of the category. But it’s not until the generative AI back in late 2022 to now, more agentic AI that we’ve seen a massive expansion of this category. Where there are a lot of different vendors in this now emerging category which is probably anywhere between 5 to 10x bigger than traditional automation because there are a lot of additional use cases. And the way I would kind of articulate the big difference is even with RPA or Intelligent Automation there was a limit to how much you could automate a process because think of things like unstructured content, or making decisions on ah, let’s say a product replacement process where you have to look at the customer’s contract, you’ve got to look at what is the right replacement product, is it available, is it available in the region, what’s the pricing so what are the price discounts you got to give based on the customer? There’s a lot of various types of information including looking at the product catalog that you have to look at. But historically it’s just hard for process automation to do it because you have to look at a lot of different unstructured content and make a decision. Now with Agent Take AI you can do those kinds of use cases and there are other use cases around prior authorization and AML and many others. So what Agent TK has done is it’s expanded the aperture or the potential of what you can automate. So if only you could automate maybe 20 to 30% of your processes now you can go all the way up to 50, 60, 70%. So that’s a huge time expansion. But there’s a huge capability expansion for our customers. And that’s been our focus ever since late 22 which we’ve been all in on. First generative AI, then agent AI, but really focusing on creating first kind of AI assisted experiences and then AI native experiences where a lot of our new products especially in the last two years are built ground up with AI and that’s really enabling customers to drive massive expansion in what they can automate.

>> Craig Gould: Previously I worked in financial services in mortgage originations at one point. And so in that world you have a customer facing person, a processor and an underwriter. Right. And you know it sounds like you guys, one of the places where you guys really found an ability to provide a lot of improved efficiency is in that processing because it’s, I used to watch these people do their jobs and it’s highly repetitive tasks but there’s a lot of moving pieces and there’s lots of judgment that has to go on. Right. Can you talk about your guys involvement in that sort of process in financial services?

>> Adi Kurungati: Yeah, so mortgage origination, loan origination, those are typical processes within financial services. so the way it would work is so we have this kind of Zenjun Mozart orchestrator essentially which allows you to kind of automate a process from retail operations all the way to kind of the mortgage team, so to speak, Legal because legal is involved and then the ops team, as you can imagine a bank, any of these banks are, they got multiple of these functional areas and but our process typically flows across and historically these departments areas have been big silos. So typically in the old world the way it would work is you get an email, you get a portal where a mortgage input comes in a Bunch of documents that come in and someone has to manually review everything and then port it over to another system. Then they go to legal aspects of looking at the mortgage, looking at the banking statement, looking at all the other credit, the credit scorecard and various other information and then put it over to retail operations. It’s like, you know, everything. There are a lot, a lot of manual steps. So the two things that we, that we work with our customers. One is how do you seamlessly connect across these departments? Because ultimately as a process automation company, our biggest value is that we connect seamlessly into our customer’s IT estate and, and kind of unlock those silos that are typically very common, these large Fortune 100 organizations. So that’s one. The second is for example, if you get a bunch of mortgage documents or get a bunch of bank statements, W2s and financial statements, how do you instead of a human reading through each and every one of them and then figuring out what is the right loan for, for this consumer. Now with agents as part of that deterministic process, you can consume a lot of the unstructured data, compare it to standards that your company has in terms of what are the different mortgage rates based on a customer profile and then provide suggestions to kind of that broker in terms of what are the various APRs that you could provide the customer. Again, ah, based on the financial statement. Ultimately it’s still up to the broker to make that decision on what, what to provide, you know, because that is ultimately human to human interaction. But a lot of the manual, figuring out what is right, kind of the mortgage rate, looking at the various documents, you know, comparing to your company standards and you know, past behavior of other customers in a simple, in a similar bracket or all that can be done by agent. And that’s the biggest unlock I would say for customers. Because as an example, one of our customers, it’s a large bank based on the east coast and they were trying to do loan approvals for an automotive customer. And so basically they’re processing lease agreements and loans on leases and loans on cars. And historically it took them seven to 10 days to process that approval because again it’s a very disconnected large organization. They were able to go from seven to eight days to six hours. And because of that shift they won that deal despite being in the smaller of the banks against bigger players. And not only were they able to win that deal, but they were able to process you know, 40,000 more loans than they could before. So that’s kind of an unlock not only on obviously the revenue side but also on the operations side because you know you’re obviously saving time and money, by basically building out these processes.

>> Craig Gould: I’ve sat there and I know that the speed of response is is a huge unlock in terms of knowing that if the process takes too long your customer is going to go somewhere else because in the end it can feel very much like a commodity. And so you’re, you know, your ability to, to close that, that process sooner than later and it allows frees up more time to put more sales into the funnel. And so your, your ROI has got to be a real open and shut case there. I mean what do you guys typically see for an roi?

>> Adi Kurungati: So that’s the biggest impact, right? So we talking about key business KPIs. It’s not meant to, it’s only about employee and productivity which is nice but it’s really about business roi. The typical ROI we look at our customers look at, there’s a lot of big focus on operational productivity which is cash flow. Cash flow is the number one. So that’s why our primary buyer typically, typically happens to be the CFO or chief operating officer. You know, think about supply chain operations. There’s a lot of how do I reduce bad debt, how do I improve accounts receivable, how do I, it is improved cash flow. So that’s one big use case. Regulatory and compliance is another use case. So think about tax, laws, various states, various locations, how do you ensure that you don’t have overpayment or underpayment. Overpayment you’re not going to get money back from the government and underpayment you’re going to get fined by the government. So that’s another use case including anti fraud which is AML and banking. There are lots of other use cases around regulatory and compliance. And more and more we’re seeing especially again with the unlock of agentic AI within our within process automation, more revenue generating use cases like the loan origination was a perfect example where you’re not only improving operating costs but you’re also winning business. And so those are revenue generating business. So those are the clear roi. We’re talking about millions of dollars for customers in any given year. So that’s an unlock for them because in addition to freeing up humans to, or employees to do better work in the customer service in the loan origination example. It also frees them up to do what they do best, which is engage their customers. I mean nobody wants to be doing and look and try to figure out what’s the right loan approval. What they want to be doing is creating the right relationship with the customer so that customer sticks with them versus going to another competitor. so I think that’s, I would say one of the bigger unlocks for our customers.

>> Craig Gould: How off the shelf is automation anywhere? Is professional services building one off deployments or are you able to learn and further develop your products based on your customer deployments?

>> Adi Kurungati: So it’s a combination of the two. at the core we are a process automation platform which is our agent process automation platform. We have solutions built on top of it as well. So we have a solution for finance, operations, service operations, explanation of benefits which is very common in healthcare. And the underwriting. There are lots of different solutions that are built out. so that accelerates you know, obviously our customers time to market for, for many of the mid size and you know, maybe the low end of the enterprise, it’s kind of, they can build it on their own. it’s, it’s local now we have a lot of AI assisted so you don’t even need to know the product and you can basically put in a prompt and create the workflow for you and you have the right governance and can define all that stuff. we also have citizen developers who can use or business users essentially who can use our products is built with that in mind. But for the more complicated large enterprises, especially when you have to connect to whole lot of systems, legacy systems, modern cloud based systems, you got to drive these processes across departments which by the very nature tend to be siloed. So that’s where you’ll have professional services. and typically it’s not necessarily our professional services. We typically work with GSI’s like Accenture, PwC, Deloitte where you’ve got to work on okay, what is the process and then how do we tie it in? Because often some of the data is behind a customer’s firewall. And that’s another big unlock for our customers is because coming from Salesforce, I always believed that everything was cloud. Like I didn’t even have this concept that there’s this thing called on prem or even private cloud for that matter. And you realize when you are essentially automating mission critical processes, touching mission critical applications and systems like healthcare systems and financial systems, a lot of that is behind a firewall. And so you kind of need to have this more hybrid architecture where you can orchestrate these business processes, but where the data can reside in various locations, some in the cloud, a lot of it behind the firewall. so that’s where that adds additional complexity, which is where whether it’s the IT team at the customer side or professional services that are typically involved in those scenarios.

>> Craig Gould: If you could help put me in the shoes of a CIO that’s trying to solve this problem and maybe the answer is different based on the size of the organization, but there’s probably a decision on do I have the manpower to build it or buy it or something in between. What are the realities of that decision making process?

>> Adi Kurungati: So typically. So, when I speak to CIOs, let’s say Fortune 500 for now, it could differ. It’ll obviously vary based on the size of the company. A couple of things the CIO’s looking for. One is they have a customer. Their customer are the lines of business leaders. The lines of business leaders ultimately are looking for specific business KPIs that they need to drive, whether it’s NPS and customer support, a better customer experience, maybe, a better supply chain operation, whatever might be the use case. Right. So it’s not, in some ways they are serving a line of business. And when they’re looking at vendors like us, the first pillar is for most CIOs, especially Fortune 100, they have a complex ITS state. There are lots of different applications, lots of different systems. As I just mentioned, some are cloud, some are on prem, their regulatory and compliance needs across. They need to have clear auditability and governance. Especially with AI agents. What are the agents doing? because these are probabilistic in nature, sometimes they might change in how they behave. So how do I have that level of visibility? But the number one question I typically get is how easy do you make it easy to connect into my existing ITS state? Because too many vendors try to go in and say, just come to us and we will solve all ills. Our POV is different. The way we present to customers is different. Because ultimately that’s our process automation, heritage, so to speak, where our entire goal is to help the customer automate processes across applications and systems. So our entire goal is to make it easy to connect into that existing ITS day. So that’s really important for CIOs. Come to me where you can connect it to my tsa. Don’t try to create another snowflake, another wall garden. So to speak. The second is the governance around it. So how do I have real time observability and governance especially with agent A.I. can I see? Because often it’s the line of business, these processes running across lines of business. You might have a hub and spoke model where you have central IT managing but you also have IT teams within the lines of businesses managing their business outcomes. But do I have a centralized observability and governance of those processes and whether when they’re touching these critical systems and are those agents who are acting on those critical systems. Can I take is there resiliency built into the system? Because things can’t go down. There has to be 99.95% resiliency. and then finally the openness of the platform because every company is going to use different models, they’re going to use different vendors. So one of the benefits that we provide is well you can build a process on our platform and you build agents also on our platform. We allow you to connect to any model, any rack service, even models that are completely on prem like open source models and you can even use agents built on other platforms. And that’s the biggest unlock for customers is because they don’t have to pick. You might have using Agent Force and Salesforce just as example. You might create a case summarization or case or case classification agent. Well you want to reuse that. You don’t want to create yet another case classification agent, another spot. Or you might use crew AI ah to create an agent on crew so you can reuse that as part of the automation anywhere process which is this entire concept of interoperability and openness is a big unlock again for our customers. Regarding the core aspect of build versus buy, I think it’s always going to be a mix. They’re going to be pieces where you don’t only buy into it. Like I think very rarely do I see a CIO say I’m going to build entire process automation engine. Because that’s a lot of work. That’s like you’re becoming a product company. But I want to necessarily I might build an agent using open source or I might use different vendors to build different types of agents. And that’s okay. You don’t need to actually only use one vendor to build an agent. And because there’s a you know what you build in Service Star and Salesforce is going to be different from what you build at SAP is going to be different from what you build in automation area. And that’s all fine as long as they all work together. And that’s kind of the interoperability piece is really important.

>> Craig Gould: Can you talk a little bit about trust? I’ve been listening to you know, CISOs talk about governance, managing identities. You know, some of them suggest that your new agents you need to kind of treat like interns and then as they demonstrate their capabilities, give them more responsibility based on the trust. Can you talk about that kind of thesis and how that does or doesn’t align with your guys methodology at Automation anywhere?

>> Adi Kurungati: Yeah, I actually agree with them in, in spirit that we’re still early in production grade deployments. We have I’d say 2000 production grade deployments. That’s 2000. We had like we had close to 500 million processes running. So still early in adoption we have about, I’d say about 5 million agent executions in production. so being an intern is not a bad thing. You know, an intern from college, you are learning and you’re hungry and all that stuff. we’re also working, there’s also this concept of agent identity. What is agent identity? Is it because often agents are acting on behalf of a human right? Because like in that loan origination it’s acting on behalf of that broker to figure out what is the right rate. But do you give the agent full access like you would give a human? What we are seeing is more, it has to be a, it’s often a limited access till you trust the agent because sometimes the agent can start doing things. You need to give it a little bit of a, clear space of where to act. But you don’t want to give it too much too much freedom to act because it can also go off the rails. so there’s an aspect of agent identity, so giving it clear context is super important so that it only can act within that context. You can’t just kind of do stuff on its own and then additional controls like what tools you give access to this agent. Especially in the enterprise setting, we have tools now like Open Claw, and you have Dispatch from Anthropic coming out. So you’re getting into a point where you have agents that can even create tools based on the goals. You don’t even need to give it tools. It can create tools and these tools could access file systems on your computer. So there is I can imagine for a CISO these are areas of concern because these agents can do a lot of bad things. They can do a lot of good things, but they can also do a lot of bad things. So what we’ve seen with our customers specifically because again we touch these mission critical processes, is definitely let me have a lot of auditability trust. There’s a heavy focus on agent eval. So evaluating an agent behavior at design time but also at runtime so that you can see is there a drift in behavior. You expected a certain behavior, but the behavior has not changed. Giving it access and credentials which are more limited versus what a human would have. And as you get more comfortable, yes, you can give it more access, but you still have the audit, the governance, the real time observability and the ability to retrace your steps if anything goes wrong. And we’re still seeing primarily human in the loop where an agent acts. But the ultimate decision is for the human to make. I think we’re still at that phase. Some of our customers have more simpler tasks where let’s say they want to do a summarization and then update that, let’s say a case summarization and want to update the outcome of that case summarization into a case in Salesforce. Those things they might say okay, let the agent do it. But specifically if agent needs to respond to an end customer or needs to update a financial system is definitely a human in the loop because those, those actions can be, there can be a lot of harm if the wrong action is taken there.

>> Craig Gould: Do some organizations go as far as to identify the agents as digital employees and try to track whether or not they are staying within the scope of best practices in the organization?

>> Adi Kurungati: Yes, I think, now I don’t think it’s, there’s a standard yet for agent identity. In fact, we’ve been talking to some of the security vendors and we’re all thinking what is that agent identity? Just like, you know, this entire concept of credential walls and you can have access based on this, credential walls and stuff like that. So is there a similar concept for agents? And that’s not like being fully defined. It’s being defined as we speak. But we definitely see even today where you’re giving certain identity though it might be using a similar credential wall concept, to agents. So you have not the full access, but you can identify that the agent actually did this work and you have then the full audit history on how the agent reasoned, planned and executed on its goal. And that’s fully visible to the IT admin so they can take a look at it, and the other thing with the Open Telemetry standard, what we’ve enabled is not only looking at our agents, agents, builder automation anywhere, but also agents that you might be using from other Systems. So using OpenTelemetry you have full view of how agents are acting across pretty much any provider. And that’s a big unlock for you know, IT leaders, even CISOs, because ultimately they have to look at that holistically because it’s not only looking at agent behavior, it’s also looking at agent behavior in the context of that process. Right. because you can have a summarization agent or a case summarization agent, but if you do it in a claims adjudication process versus another process, you need to know that context so you have that full view, based on again a defined agent identity. I think what we’ll find shortly is a more standardized way to define agent identity. And that’s something we’re working on with

>> Craig Gould: certain security vendors and identity providers in the larger organizations. Where are CIOs getting it right and where are they getting it wrong? If they’re trying to integrate the, this automation and they’re trying to automate processes, integrate agentic AI to improve these line KPI’s, where has there been a little too much exuberance and maybe going too fast or I don’t want to put any words in your mouth. Where, where are our CIOs getting it right and where are they getting it wrong?

>> Adi Kurungati: So I, maybe I’ll do a couple of scenarios and then we can go a little bit deeper. where I’ve seen this work best or agent tech automation work best is where there’s a clear outcome you’re going after. There are a lot of scenarios where hey the tech is new, it’s cool, I will figure out what to build. The copilot is a perfect example. What happens. Ultimately the ones who have to use it are your business teams because most processes cross across business teams unless it solves a problem, or driving outcome. what typically happens is you try to deploy these agents, or these postal productivity type processes and nobody uses it. And so then you go on shelf there. And part of the thing about agents is the more user the better it gets because it’s learning from past actions. And that’s kind of something we’ve also built into our product. So outcome driven critical, starting with clear, that’s why we love operations and finance with clear roi because that really then you can build on that building human loop Scenarios from the get go. Not assuming that the agent will be able to execute on its own even if it can and if you’re pitched by different vendors that you should just let the agent do whatever it wants to do. I would not do that. I would have always you in the loop scenarios and clear centralized governance to look at what’s happening, what is the audit trail, are you having agent eval so kind of building that the centralized governance architecture before you do mass deployment. so typically the best deployments are a collaboration with business and IT where they’ve created a well defined governance structure. What will central IT own, what will the business IT own and what are the three or four outcomes we want to drive towards? That’s what they’ll drive. They’ll show the ROI for IT and then scale out. Those are the best. And the ones that don’t work as well is where you just put a bunch of copilots in front of different teams and hey, just use it. And that’s our AI story. That doesn’t necessarily work in my opinion.

>> Craig Gould: So there’s some AI washing there. There’s public relations speak that we’re integrating AI, but it’s really just, just lip service versus actually having efficiencies that have been identified that you, you want to achieve.

>> Adi Kurungati: And one of the things we do, pretty much for all our customers and frankly even prospects, we open it up to anyone. We have hands, on sessions where they can actually build specific outcomes. There’s nothing like building like you know, I build using cloud code and I’m using various other tools out there because as you said there’s a lot of noise out there in the market but you don’t know what applies to you till you actually get your hands on it. And so that’s one of the things, I think best CIOs know they need to get their teams on it to actually try these different tools and try to understand what actually works versus what, what is more AI washing or you know, marketing fluff out there.

>> Craig Gould: So how do you guys enable that with, with your customers?

>> Adi Kurungati: We have a Pathfinder community, so we do, I would say every month, sometimes every other month, virtual sessions where we can have three, four thousand folks, they can also do it on their own. They don’t have to actually attend the session, they can do it offline, but they do entire build session on a set of scenarios. And so they’re building a process with agents and they’re building an agent and figuring out the governance and all the other things so that they get comfortable with it. It’s not theory and I think that nothing goes. That’s the best way to no loan.

>> Craig Gould: So do you guys have like sandboxes for different industries where they can kind of see here’s a scenario within my

>> Adi Kurungati: industry and I think we typically first start with a couple of well defined scenarios, a general sandbox and everybody has access to it. We also have sandboxes like specific use cases for financial services, manufacturing, healthcare. we don’t necessarily do it for every industry, but those are the three

>> Craig Gould: main once we do, there’s a lot of noise out there about what agentic AI is capable of. I want to make sure that I stay focused on the key differentiators that are making my company profitable. And I don’t want to be distracted, but I also don’t want to get left behind. I don’t want to be continuing to make carriages while Henry Ford’s making automobiles. Right. And so how can I make the wisest decisions at this state of where the world is in relation to automation?

>> Adi Kurungati: Without a doubt, this market, the technology is changing rapidly. while there might be a little bit of hype, there’s without a doubt the technology has changed, the outcomes that are possible have changed and it’s only changed faster. I was talking about open cloud. Now you’re going to have open cloud for the enterprise, you’re going to have other use cases come in. That’s not hype. The technology itself is not hype. The application of it and the maturity of it definitely will take time. So I think it behoves m the CIO to be on, on top of that game to understand what actually applies to their business. And so again going back to hey, what are the core outcomes and for those core outcomes, how can I use this technology? I’m not using technology for the sake of technology. I’m using the technology to drive those outcomes. In my own organization. We drive a lot of experimentation because you’re constantly experimenting with different tools, open source, vendors, whatever it might be, even consumer tools are figuring out what are those consumer tools we could incorporate into the enterprise with the enterprise setting, with enterprise structure around it. So rapid experimentation or constant experimentation within your teams. I would even creating a kind of AI team if you don’t have AI. Most OCIs now really have very strong AI and data teams where they’re also rapidly experimenting and it’s a collaboration between them and maybe DevOps and IT ops, because they need to know, one on the engineering side, one on the more AI side because both have to go hand in hand. Also instituting a governance, council, to make sure that ultimately, as we are rapidly experimenting and we now are ready to incorporate some of this agent Aki into a core business workflow. You have the right. The governance Council reviews it. So having that structure in place is really important. That way you have enough sort of checks and balances to make sure you don’t do the wrong thing, out there for your business. But so it’s kind of the balance of, on the one end you can do nothing and say, you know, this is all a, fad and it’ll go away. And just like, you know, there have been fads in the past on the other end, let’s go all in, we got agents all over and let everybody go. Helter scheduler. I think for me it’s more nuanced where you need to have the governance, you need to have the experimentation, but you need to constantly think about how are you going to use AI, broadly, but agentic AI specifically to drive those business outcomes for your core customer, which is the line of business.

>> Craig Gould: Right.

>> Adi Kurungati: And so it’s a little bit more nuanced, but I think that’s where the best CIOs, that’s what they’re focusing on.

>> Craig Gould: You work in a job whose your title didn’t even exist 10, 15 years ago. You know, if you were giving some advice to someone that is in the middle of their career and they’re trying to get to the next level, they’re trying to, you know, you know, to navigate their career. Given the fact that you have navigated to a destination that you couldn’t have even foreseen, what advice can you give someone on that type of navigation? In a world that’s kind of constantly changing, how do you identify opportunities and how do you take advantage of them?

>> Adi Kurungati: It’s a good question. So my background is, and I, used to be a developer way back and then a product. I’ve come from the product side of things, before taking over in this role of Chief AI and Development Officer. I think the couple of things remain the same, which is thinking about what differentiated outcomes you’re going to drive and how do you quarterback building those differentiated outcomes and constantly listening to the customer. But also not only the known but met but unmet needs, but also some of the feedback from customers, like looking at what they try to do versus what they say right, because sometimes some of the products you’re building, m they might be solves for the actual outcome they’re looking for. But the customer may not tell you that this is what we exactly want. So having that kind of customer lens and outcome lens I think is really good. The other is if I look at one of the things that helped me was the hustle, back in the Salesforce days there’s a hustle to build product and kind of cut through the noise and be a little bit more aggressive. With my team together we were like a startup within a big company because Salesforce is a huge company. And the way we were able to build, deliver and build $1.5 billion business line over over seven years was that we were constantly experimenting and pushing the envelope. I think for any leader out there, AI is a perfect opportunity to a constantly be experimenting. But experimenting with the construct of how do you use AI and agentic AI to deliver outcomes faster with higher value and differentiated value. So if you can do that and you can do it as part of a team, you want to get hands on. I think there’s no shortcut to getting hands on, to using AI to understand what’s real. Constantly experimenting. You can’t only tell your team hey, go experiment and let me know how it is. You have to also be experimenting because that way you get a sense of what can work, what can’t and you can have an intelligent conversation with the rest of your team. So constantly experiment, have a set of folks around you who also have that mindset. Because often we all grow in our careers based on who we follow and the teams we’re part of. It’s not only about the individuals, especially in product and technology as part of the team that you are. So I think all those can help you navigate and kind of grow to the next level. And always been learning mode. I think that for me has been something that I’ve been very focused on. when I joined Salesforce I didn’t know what leads were, leads and opportunities were. When I joined automation anywhere frankly I didn’t know what RPA was. I had to kind of look it up. and AI or agentic AI, it’s been all a learning for the last three odd years. Right. None of this was known. Not that I studied transformers, back in school, any of that stuff. So it’s just kind of having that open mind to learn, to experiment and fail fast sometimes, but also be part of that team that together you’re building something and delivering value to your customer. If you have that mindset, I think opportunities open up because that’s what everybody wants. That’s the kind of leaders I want who are constantly learning how to kind of hustling and building amazing products and solutions, that provide differentiated value to our customers.

>> Craig Gould: I love it. Curious hustlers. That’s that’s the secret. Well Adi, I really appreciate your time today. you know, this has been so informative and just so timely. If folks wanted to keep track of you or automation anywhere the developments or if they wanted to reach out to the product or sales team, how’s the best way to keep track of you guys or touch, base?

>> Adi Kurungati: So we have a very active Pathfinder community. We’re very active there. We have this podcast called the Agentic Edge. here. There’s a lot of content there. you can obviously reach out to us on LinkedIn. There’s a lot of on the website as well. But those are typically the best ways to engage with us. And we’d love for your listeners to even participate in our Pathfinder community events because that’s where you also meet other customers, and get, as I mentioned earlier, hands on with some of the technology that we’ve been launching.

>> Craig Gould: Wonderful. Well again Adi, thank you for your time today.

>> Adi Kurungati: Thank you Craig, appreciate it.