UiPath Fusion 2025 | Maxim Ioffe, Wesco

Clips
More from UiPath Fusion 2025

Maxim Ioffe

Global Intelligent Automation Leader

Wesco

play_circle_outline Introduction to UiPath’s Fusion Conference, previously known as Forward.

play_circle_outline Importance of process over technology for delivering business value.

play_circle_outline Enhancing Automation: Transitioning from RPA to GenAI Orchestration While Upholding Standard Operating Procedures

play_circle_outline UiPath's strategy to be the orchestrator amid competing SaaS solutions.

play_circle_outline Barriers such as unclear processes and data limitations hindering automation efforts.

Info
Transcript

Maxim Ioffe, Wesco

Maxim Ioffe

Global Intelligent Automation Leader Wesco

In this UiPath Fusion 2025 segment from Las Vegas, theCUBE’s Dave Vellante sits down with Maxim Ioffe, global intelligent automation leader at Wesco, to unpack why “the process is the hero” in the agentic era. Ioffe shares why this conference’s sharp focus on ROI resonates with practitioners and how orchestration is evolving from stitched-together RPA to coordinated systems of bots, agents and humans. He explains Wesco’s best-of-breed strategy, the appeal of UiPath’s “Switzerland of automation” posture and the practicality of swapping technology “blocks” with... Read more

explore Keep Exploring

What is the current event being discussed and where is it taking place? add

What is the relationship between process and value in a business context? add

What are the considerations and challenges in automating business processes while maintaining existing standard operating procedures? add

What is the perspective on the future direction and strategy for UiPath in the context of automation and orchestration? add

What are the main challenges faced when trying to automate a business process? add

bolt Powered by CUBE AI

Maxim Ioffe, Wesco

search

Dave Vellante

>> Hi, everybody, welcome back to Las Vegas. We're here at the Wynn. This is day two. We're wrapping up day two of UiPath's Fusion Conference. It used to be called Forward. They did seven Forwards, dropped the name, they transitioned to Fusion, which is great, because we're fusing the past with the future, agentic, bringing new levels of opportunity in AI. My name is Dave Vellante, and we're here with Max Ioffe, who is a CUBE alum, and global intelligent automation leader at Wesco. Good to see you again, Max. Thanks for coming back on.

Maxim Ioffe

>> Thank you, Dave. Thanks for having me. Excited to be here.

Dave Vellante

>> How is Fusion going for you? You've been to many of these customer conferences, what do you think of this one?

Maxim Ioffe

>> I loved it. I think that's the first conference I've been to where ROI was at top of the stage. Bringing the ROI into AI is something that is really, to me, it's the right to area of focus. And it's one thing to put it on a banner, it's another thing to try to put it into just about every conversation, just about every talk. A lot of keynotes focused on that. It's not the hammer looking for the nail, it's more about how do we do it in a profitable way.

Dave Vellante

>> One of the things that Daniel said in his keynotes, which really resonated with me, is he said, "The process is the hero." And I think about that, because a lot of times we talk about data gravity, data has gravity, and people don't like to move data, but process also has gravity, and it's really where, you think about it, in and of itself, it doesn't deliver value. What delivers value is how it supports the process, how the process supports the business, increasing revenue or cutting costs. So how did that resonate with you? What do you take away from that?

Maxim Ioffe

>> No, it resonates really well. To your point, the technologies, of course, technology is not necessarily the value. The value is in doing what business needs us to do. We, as technologists, don't tell the business, "This is how you should run." We come into business, in my case, a distribution company, a wholesale distributor, and say, "What can we do to help us to work well with our partners, to service our customers, and hopefully make some profit out of it?" Well, that's all about the process. It's all about the individual steps in that big thing that we are doing.

Dave Vellante

>> How do you think about... I mean, you've got a ton of experience in RPA, automation, you run the center of excellence for your company, and have for a number of years. And now with GenAI, we're this new era of possibility. I mean, it was almost like we were trying to force previous eras of automation into this new era, and then GenAI comes along, it's like, "Okay, now it's no longer aspirational. The technology is going to support it, but now there's a human element to it." So how do you think about that journey from point products, like RPA, stitched together sort of various tools to create the beginnings of a platform, to which now is we're entering a platform in a agentic era?

Maxim Ioffe

>> Yes. I mean, we are slowly segueing into orchestration in a way, right? But fundamentally, if you think about the business experience, we want to automate the process. We look at the process, we see some tasks that are pretty straightforward, their roles-based, there's no need to change anything. We can still automate them. There are some points of decision-making. Well, technology allows us to automate some of that decisioning, and that makes it a lot easier to automate and deliver business the solution. But at the same time, we, as humans, are creatures of habits, creatures of our standard operating procedures, creatures of "This is how we want the business to be done." And we want to be careful to preserve that throughout the automation, which means that as we look at the use cases, we still look at it through the lens of "What do we want to achieve? How business wants to do things, now let's see if we can do it with technology." A lot of times the answer is yes, and a lot of times it's agentic, robotic, and human in the loop, and each one of them can shine and deliver the best with what they're capable of delivering.

Dave Vellante

>> So when you think about things like process mining, and the like, we were looking for opportunities to potentially automate, and now it feels like, and I'd love your thought thoughts on this, it feels like we're entering the era of beginning days. But the North Star is that the operational model will be that the process is built on the fly, if you will, to serve the particular need, and according to objectives and constraints and metrics, and whatever those are, that it'll be optimized for that particular task. It's like every process actually could be a beneficial snowflake, whereas before, the last thing you wanted to do is to have a snowflake for every situation. But now, it seems like within the next several years, we'll be able to get there, where that individualized, that personalized process is actually, that snowflake is actually an advantage, if that makes sense. What are your thoughts on that?

Maxim Ioffe

>> Yeah, it does. And I think about it as a snowball versus snowflake. When you look at it from the macro level, yeah, it is a snowball. It's a big process. It has a lot of different things that need to happen, hopefully, in the most automated, the most agentic, the most smart way possible. And within that snowball, as you start zooming in, you start seeing, "Oh, we're still doing tasks. We're still entering orders." We're still chasing our suppliers saying, "Hey, when is my product coming?" EDI has been around 50 years, it doesn't mean that it can answer that question in an effective way and it scale. APIs has been around for decades, it doesn't mean they can answer that question. Sometimes good old web-scraping is what takes you there. And if you think about UiPath, if you think about Maestro Orchestrator, it's exactly that. To me, it's all about starting from that big snowball of a process, zoom in, and you start seeing the blocks and decision points. Zoom in some more and some more, and eventually you see a dozen of bots, and a couple of agents, whatever the ratio is, but that's what does it, and I don't see that changing. We just can do more, and we can visualize it a lot better. And frankly, I think from a business user standpoint, we just see a snowball. You don't need to worry about all those individual elements that build it.

Dave Vellante

>> That said, how do you see the software stack changing? Think about the move to SaaS, it changed. The technology model became cloud, the operational model became standardized templates from snowflakes in the previous on-prem, and then the business model, the pricing model completely changed. And it seems like we're going through a similar transition now. We call it, instead of services as software, some people, we picked up on this... Sorry, instead of software as service.

Maxim Ioffe

>> Software as a service, yes.

Dave Vellante

>> Services as software, where those services are productivity drivers that are delivered through software with new services. Before it was, whether it was HR, or service management, or CRM, now it's new services delivered through software. So how do you see that? How do you see the software stack changing as a result, and what role does orchestration play in that?

Maxim Ioffe

>> Great question. And I think we have to live through that over the decade that we are trying to consolidate various ERPs, and decide where we want to be. And eventually, we made a decision that best in breed is where we're going to go. It's not about that big analytic software package that does it all at once, it's more about who is the best at accounts payable, who is the best at supply chain management, who is the best at whatever business function you're looking at. And all of a sudden, you have this proliferation, where you cannot even count how many software packages you need to work with. They all need to talk to each other, and that takes us back to the orchestration. Half of it is exchanging the data, but a bigger part of it, back to where we started, is the processes the king, and exchanging the process and seamlessly integrating all this different stack into a unified process, that a business user doesn't need to worry about what's behind it. And frankly, best in class, I think Daniel mentioned the Switzerland of automation, like bring your own model, bring your own solution, bring your own anything. This continues to grow, and that's where the orchestration becomes awesome that you bring it. We need to be almost instant at unplug this and plug in this, and guess what, that snowball of a process stays the same. It doesn't even notice that something changes, except that the process has done faster, better, whatever the case might be.

Dave Vellante

>> I look at it similarly. We use the analogy of Lego blocks that can fit together and be formed to do whatever you need it to do that at that point. But in the context of Switzerland, so I'd love your thoughts on UiPath's strategy in that regard, because the process is the hero, and the SaaS vendors have the process logic. They have the metadata, and the data associated with the process, and they're going to try to hang onto it. So UiPath's strategy is to connect to those, form partnerships, and be the orchestrator. I feel as though there are others that are maybe in their SaaS domain, also want to... You know, "That's our task to orchestrate," and so that creates a little bit of tension. What breaks that? Is it the customer demand, that says, "No, we want no lock-in, we want the horizontal capability," versus the SaaS vendor saying, "No, no, use our orchestrator?" How do you see that? Or there's like an orchestrator of orchestrators, how do you see that playing out?

Maxim Ioffe

>> I mean, every customer is different. Every company is different. Every software provider is different. Of course, Salesforce would want you to be locked in Agentforce, and stick there, and service now wants you to be locked in their ecosystem. And we, as a customer, need to make a choice, is that where we want to be, or do we want to be able to bring our own model. Look at the AI market, how quickly you go from GPT-4 to GPT-5, and then Google does the next version of their product, and somebody else does this. It's a decision-making. And to me, what UiPath is trying to say, at least the way I understand it, is bring the best in class, substitute the best in class. Back to your Lego block analogy, pull out this red brick, put in the blue brick, it will still work. To me, that looks very attractive. And if we, as you said, orchestrator of orchestration, I think that's where it's headed, and I think that's a very smart move for UiPath to take that role. At the end of the day, the automation plays that narrow gap between poor software capabilities and purpose-built software tools that can do it, and it's almost like the orchestration is what goes way across that, as opposite to sticking the automation into that narrow thing. So I think it's a great plan. The fact that they have Ui automation, it basically means that if I have an API, amazing, I will integrate with APIs. If I don't, I still have my robots that can act like an API. And we have that ability to fuse things together that are not fusible on the surface.

Dave Vellante

>> So you see building on top of existing automations as an advantage for your company, is that a fair assumption?

Maxim Ioffe

>> Absolutely. I have many hundreds of automations, I look at each one of the automation as its own tool for a possible agents, or think of those as deterministic agents. So if I have a deterministic process, I have a deterministic agent to do the work. If it's this non-deterministic goal-driven process, I have agents to do the work. And where I need to insert humans, I need to insert humans. UiPath makes it very easy. And it also could be a journey, where I start it as a human-driven process, think of a Maestro flowchart, and every block says "Human, human, human, human." All of a sudden, this block can be a bot, this one can be an agent, this one can be a combination of the two. And we slowly drift toward that big snowball where everything is done. The way we want it to be was humans only approving or making the key decisions that we really don't want to automate regardless if we can

Dave Vellante

>> From a customer perspective, and as somebody who's an automation expert and has seen it all, what are the major barriers, blockers, to use that term, that if they were removed, would make your life a lot easier?

Maxim Ioffe

>> It's a good question. I think the biggest blockers have nothing to do with the software. It's a business reality. We need to define the process. In order to automate the process, we need to agree on the process in order to automate the process. A lot of times, you even say, "Well, you know, agentic is great. Can you build an agent that does something?" Okay, we can do it, what's the good outcome of that agent? We don't know. Do we have the data to do this? We don't. Those are the barriers that will be solved over time, and it's making the journey a little bit less straightforward than "I'm going to build the agent today." It might be "I'm going to build a process that collects the data today so I can have enough data to build the agent tomorrow, next week, next month." But it's a journey, and the journey is going to be walked just like we did with RPA.

Dave Vellante

>> So that's out of the control, generally, of a software product company.

Maxim Ioffe

>> Mm-hmm.

Dave Vellante

>> Who are the types of companies that you want to see UiPath partner with, you don't have to name names, or you can if you want, that can help remove those blockers?

Maxim Ioffe

>> That's a good question. I think UiPath have built a really nice ecosystem, where they are bringing a lot of things, right? And we haven't talked about security, but obviously, it's not a blocker, but it's a consideration, so partnering with companies that can help to make sure that the trust layer is as secure, as trusted as possible. With data integrations, the clarity and the visibility of the process, all that needs to be happening. I think your hypothesis on the right path, no pun there, to bring all those companies together into the ecosystem. The fact that they're taking the approach of "Bring us your model, bring us your LLM, bring us your agent, bring us your agentic framework," is telling me that they are thinking about it the right way. They may not necessarily even need to partner, we can just integrate it together. I think that's a great idea to play in the orchestration layer.

Dave Vellante

>> Well, since you brought up security is the, we haven't talked much about security. We were here two weeks ago, at the CrowdStrike show, that's all we talked about, is security, and it's a very important topic. What is the security model for things like role-based access control, or making sure LLMs don't leak, or encrypting queries, returns, data, data protection, et cetera? What is the security model in the automation world? How much of that is the responsibility of UiPath? How much are partners, whether it's identity players, or certainly, big cloud players, et cetera?

Maxim Ioffe

>> Good question. And I think security is a very broad word, and I'm sure CrowdStrike folks are a lot more qualified to answer that question than I am. But ultimately, it's all about organization's place and willingness to take the risks, and really let the IT security experts, like the CISOs professionals, everybody else, tell us how they want it to be, and let UiPath, to your earlier question, bring the right partners to deliver what needs to be delivered to match that security. I'm not an expert in security, I'm an automation guy. I vaguely understand the concepts. To me, the key here is your possibility to observe and talk to my security, and say, "Yes, we do have the right certificates. Yes, we do know how to protect the data in transit, at rest. We know where the data is going," and all that stuff, right?

Dave Vellante

>> Okay, so the scenario is I'm an organization, I have a SecOps team. They might use, whatever, Okta or Microsoft for identity, maybe Palo Alto or CrowdStrike, and whatever it is, and a number of tools, that's their domain. UiPath has to make sure that it plays nice within that domain, which they're selling enterprise software, so we know what that's like.

Maxim Ioffe

>> Yes.

Dave Vellante

>> You can't just show up, and say, "Here it is." "Oh, security, I'll get back to you." No, you have to be hardened, which they've done over the past many, many years.

Maxim Ioffe

>> Absolutely.

Dave Vellante

>> So they fit into that model. And then within the world of whether it's bots or agents, they're responsible for that piece of it, so it's kind of a shared responsibility model.

Maxim Ioffe

>> Right, absolutely. To your point, we went through UiPath contract renewal recently, and the biggest, and the longest amount of time that I budgeted for was security reviews. I knew it's needed, and I wanted UiPath to talk directly to our security team and make sure everybody's comfortable, which we did. And then to your point, each agent, each bot, has its own layer of security and responsibility. This is where it's on us, as an architect, on enterprise, as what is the right way of doing it, and on the security providers to make sure that that right way actually works and doesn't bring us down, like certain events done this CrowdStrike a while ago.

Dave Vellante

>> Yeah, yeah.

Maxim Ioffe

>> It was memorable in a way, that it shows how security is important, how well it controls, and can throw things out of .

Dave Vellante

>> Right, and that wasn't even a security event, it was an IT event. It was a patch. So that's-

Maxim Ioffe

>> Exactly, exactly. Right. That's the level of dependency, right?

Dave Vellante

>> Exactly. It underscores the connectiveness and the dependencies, to your word. I remember the flash crash, I don't know if you remember that, the stock market dropped unexpectedly.

Maxim Ioffe

>> Right.

Dave Vellante

>> And people couldn't explain it. They called it a glitch. Glitches means that we're so interdependent we can't explain it.

Maxim Ioffe

>> Right. And that explainability of AI needs to happen there, right, otherwise, we are in big trouble. And another turn to the security that we haven't talked about, I'm not enough, me in any area, I don't know what the right segregation of data is. So it's not only on me, but primarily on the business to tell us what is safe in their mind, what are the right levels of authority, whatever, right, segregation of duties. All that is a shared responsibility of security team, the automation team, the SMEs, and a lot of other players. It's complicated.

Dave Vellante

>> How do you feel about so-called AGI? Do you have any thoughts on that? As a technologist, I'm almost embarrassed asking the question, but because you say you're a generalist, people talk about AGI, I like the term "Enterprise AGI," meaning applying your own proprietary data and processes to LLMs, or maybe... I don't know if AGI will ever get there, what that even means. But I'm talking to proprietary advantage.

Maxim Ioffe

>> Right.

Dave Vellante

>> That's, to me, the holy grail. We wrote a piece a while back, Why Sam Altman, or Why Jamie Dimon is Sam Altman's Biggest Competitor. Jamie Dimon being a metaphor for enterprises that have proprietary data. Every enterprise has the proprietary data. Thoughts on that?

Maxim Ioffe

>> Right. Well, a couple of thoughts. One is proprietary data is definitely there. Everybody has a secret sauce, and that's what makes that enterprise survive, exist, and prosper. The question is where is it really saved? Is it saved in the databases? Is it saved in the process maps? Is it saved in employee heads? Is it saved in a corporate culture?

Dave Vellante

>> Right.

Maxim Ioffe

>> And if so, how will AGI extract that data? We are really at the point where I think all the easy data is already extracted, talking about AGI outside of the enterprise or within the enterprise, but there is so much data that is tribal knowledge that's not anywhere. It is implied. And the other factor to it is we, as humans, are extremely adaptive. Look at COVID, how a week goes by and everybody's wearing a mask and everybody figures out how it works, and we shut down schools, that somehow society didn't fall. Will AGI be able to adopt at the same speed to this whole new data that doesn't exist anywhere, but all of the sudden, what was absolutely the norm today, is absolutely inaccessible tomorrow, and vice versa? COVID is an extreme example of that, that business leads that world all the time.

Dave Vellante

>> Adaptability, that's kind of a human trait.

Maxim Ioffe

>> Adaptability, and this is back to Maestro orchestration, right?

Dave Vellante

>> Resilience, yeah.

Maxim Ioffe

>> Which block is responsible for it, and can I substitute that Lego block very quickly? I'm hoping that my S3 is an answer to it, in a way that it makes the process visual, and I know that, "Okay, this is the block that I might need to replace with a human until we reprogram AI to understand what it needs to do." Who knows, right?

Dave Vellante

>> I liked how you built that block diagram, human, human, human, human, human, and then you identify where the best candidates are for automation. It's not necessarily... Well, how do you think about that? It's not necessarily the easy ones to automate, or maybe it is, or is it the high-value ones, like Daniel was saying on Monday or Tuesday, "Be more ambitious. Go for the big wins." How do you choose where to replace the human? What's the right playbook?

Maxim Ioffe

>> I think there is a balance. And again, every company is different, every funding situation is different. How much money, to me, it's all about the ROI equation has to check out. The size of opportunity, the cost of the opportunity, and the cost of the return on the investment, they got to work out in favor of building it. And I think from the early days of RPA, we know that not everybody thinks that way. And those 95% of POCs that fail through MIT paper, I think a lot of that has to do, "Oh, that's a cool tool, let's build something." But we haven't thought of ROI, we haven't thought of technologies, we haven't thought of approvals. If it all connects, yes, you can go for the big case. But in order to think through all that, and gain the experience, maybe start small, dream big, eventually get there. To me, a big process, like that Maestro map that goes end to end, is really the thing that Daniel is going after. But again, as you zoom in, there are so many small steps we need to make, let's make some today with what we already can, get the ROI. Hopefully, it will fund the next step, and that will get us big pretty quickly. The good news is the development cycle with this technology is very short, and ability to extract the ROI doesn't mean that you have to build A to Z. You can build C to G, or C to whatever, I'm struggling with the alphabet for a sec, but you can build with smaller elements and figure it all out pretty rapidly.

Dave Vellante

>> 95%'s not necessarily a bad thing. It sounds bad, but that means there was a lot of learnings that came out of that.

Maxim Ioffe

>> Absolutely. It's a question of the cost of the learning.

Dave Vellante

>> Yes.

Maxim Ioffe

>> Going to a good school and paying a good tuition, but can you learn it with free? Can you learn from somebody else's mistakes? But I think a lot of that has to do with everybody wants to do something, but have you thought of security, have you put , have you gotten all the right data elements? Do you have the data? I mean, we get excited easily. It sounds easy to automate a process. Oh, we can automate customer service. What do they do today? Do they have the ticketing system? Where is the data sets? Oh, we have an asset up front. Well, you are in the 95%.

Dave Vellante

>> Well, Max, it was great having you on again. I really appreciate your perspective.

Maxim Ioffe

>> Likewise. Thank you so much, .

Dave Vellante

>> Thanks so much. I really appreciate it. And thank you for watching theCUBE at UiPath Fusion this year. We were super excited to be back. This is our seventh year. Check out siliconeangle.com, it's got all the news. We've got some posts up there on some of the UiPath announcements. Go to thecuberesearch.com, that's where you'll find all the in-depth research from the CUBE team. And of course, check out Thecube.net, and check out Thecubeai.com. Of course, it's our little experiment with AI, where you could just ask it questions. Ask it "What are the top-five takeaways at UiPath Fusion this year?" I'm Dave Vellante, for Rebecca Knight, and thanks for watching everybody, and we'll see you next time on theCUBE.