Why Your AI Pilot Is Failing and What It Takes to Reach Production
📄 Zusammenfassung
Dieses Video von "Why Your AI Pilot Is Failing and What It Takes to Reach Production" enthaelt keine Beschreibung und kein Transkript. Bitte das Video direkt auf YouTube aufrufen fuer mehr Informationen.
📝 Transkript
So one customer example is a financial services customer bank. Essentially reduced their loan application process for automotive loans from around 12 hours to a little under 1 hour. In all these use cases we definitely have human in the loop. Once the human approves the agent is then taking action to process it as well. I don't think the agent technology is at a point where you can get 90% plus accuracy all the time. And so there is a trust factor but for these mission critical processes that drive operational productivity or regulatory compliance or revenue impact. We believe it's a combination of determinacy and cognitive. >> AI is moving fast within the enterprise. Employees are experimenting with personal AI accounts. Teams are building custom AI apps and autonomous agents are connecting to sensitive systems. Innovation is exploding, but governance isn't keeping up. Gartner predicts that through 2026, 80% of unauthorized AI transactions will come from internal policy violations like oversharing sensitive data, not external hacks. That's a huge risk. And that's why Island, the creator of the enterprise browser, launched its new AI services. Island surrounds generative AI, AI browsers, and autonomous agents with the enterprise controls that they were never built for. Identity enforcement, data protection, auditability, and centralized policy. Now, organizations can safely use any AI app, consumer or enterprise, without corporate data loss. Teams can deploy secure permissioned agents with full audit trails. Island's publishing capability makes it easy to safely share internal AI apps across the company. And the Island AI browser brings governed multimodal AI directly into the enterprise browser employees already use. The result, a single secure foundation for AI at work. So CIOS and CISOs don't have to choose between productivity and protection. To learn how to scale AI without losing control, visit island.io. Usually I start by having guests introduce themselves to listeners uh and give uh your background so far as it's relevant but more importantly uh how you came to automation anywhere and and what exactly automation anywhere is doing. >> Yeah. Um, all right. So, nice to meet you, Craig. Uh, Adi Kuruganti. I'm the chief AI and development officer at Automation Anywhere around our product, technology, R&D teams. Um, and I've been here about four and a half years. >> Uh, before that I spent about 15 years at Salesforce.com. >> So, been in enterprise software now, I guess going on 20 odd years. uh very much in all things product for CRM now in all things automation uh big focus obviously at agentic AI and agentic and I'll talk a little bit more about automation anywhere so uh so that's my background I think just to maybe talk a little bit about automation anywhere >> we are in uh you know we've defined a category called agentic process automation which really combines what we believe is the best of deterministic like automation traditional automation with agentic AI but with the goal for customers to automate their mission critical processes things like order management prior authorization healthcare uh anti-moneyaundering and financial services so those are the kinds of you know enterprisegrade mission critical that typically span you know mult different applications different systems legacy modern as well as obviously human decisioning uh so we often touch you know financial systems healthare systems So it's really important that you get the enterprise grade governance and controls around it. We started we created this category uh before this you know um we created this category called RPA or robotic process automation I would say about I want to say 12 13 years ago that was that was a start uh and is really more focused on um automating uh repetitive tasks you know in a deterministic way. Uh and obviously that category took off. Um and you know you have a couple of our couple of the companies still remaining UiPath being one automation anywhere another one. Uh but with the last four and a half years it's really been expanding to what what we're now calling agentic process automation. So again we're in that space. So a lot of obviously focus on how do we use agentic AI to enable customers to automate mission critical processes. That's our lens. There's a lot of agent AI. Everybody's talking about agent AI for everything. We want to use it at the right time because we still believe there's a lot of value to customers to combine deterministic and agent AI. It's not one or the other. It's the combination of the two. >> It's a fascinating area right now. U I I just got Open Claw working on my computer. of >> hopefully on a on a different terminal though, different computer because uh you could get in some issues if it's your main computer. But yeah. >> Yeah. Yeah. Well, I'll take that uh under advisement because I I haven't been very careful with it. Uh but but the there there is certainly I mean that to me uh has been kind of the GPT moment for agents. I mean I I know enterprises that have been working on them for a long time. You talked about Salesforce and and they're with agent force I think the the leading one of the leading uh agent building platforms. Uh but you guys are focused not on building agents but uh but making sure they work in an enterprise. So we are we we obviously you know as part of making sure they work in enterprise we want to we also enable customers to build or what we call process agents because it has to have the process context. Um but we all you know one of our key values is having open platform because again as a process automation company we need to make it easier for customers to connect into different systems and different applications. So it can't be a wall garden, right? So we work with Salesforce, Service Now, SAP, old school terminal servers, healthcare system. So it has to work across and has to be very seamless. And so for that we obviously, you know, customers build processes on our platform. They use our agents, they using MCP, they use agents from Microsoft, from Salesforce, you know, other vendors, crew.AI, you know, open source platform. So it doesn't matter. It's ultimately up to the customer to use the right agent for the right use case and right task. Uh but that's and but really or use our platform to orchestrate the end to end process and with the governance around it. >> Yeah. Uh and there's been a lot of talk uh before we talk about uh automation anywhere's role in this enterprises been struggling with a gentech AI. I mean there's a lot of promise, a lot of mistrust and so far I haven't seen anyone although I hear about it but I haven't talked to anyone who's using it in critical business processes as you mentioned. So what kinds of critical business processes do you see Agentic AI being applied to in the enterprise? So today we have you know we have over 5,000 customers broadly but of those customers uh I'd say we had about 1,500 live deployments in production now many more uh in P world and that's one of the challenges of how do you go from PC to production and we'll talk more about that um but really typic typically it's again the three types of use cases that we see one kind of the bread and butter which is uh operational productivity So things like you know cash flow how do I reduce accounts payable the entire end to end process so I get cash uh sorry accounts receivable or get cash up front >> uh or supplier management there's certain things that you know you want to get cash faster you want to have uh and really focus on your operational productivity that's one use case second is around regulatory and compliance big use case around that so things like how do you avoid um overp payments but also underpayments ments on taxes based on different state or country uh rules because it it it varies for especially these large global organizations. Third seeing less of it but it's coming starting to come is um kind of think of it like new product line. So one customer example is a c a financial services customer bank essentially was able to reduce their loan application process for consumers like you and me for automotive loans. uh from around uh 12 hours to a little under one hour. And because of that, they were not only able to process more loans but able to win a you know basically a deal a business from one of one of the large automotive manufacturers out there and they were competing against other banks. So it's kind of operational effectiveness but also using that to win new deals. I'd say the use cases are more in the first two. So typically departments like finance operations that's where the benefit is and again it's not about using agentic AI purely like just use agents and figure out and let let them go crazy. It's about combining deterministic with these agents. >> Well, for example, contracts, it's a great use case because >> um agents when you're thinking about unstructured content like contracts using generative AI and then agents to kind of figure out what is the based on certain customer profile, what's the right loan APR you want to give this customer agent. It's a great use case for an agent to look at that customer profile, that task and figure out based on the structured content, you know, giving options on on, you know, loan APIs. Uh, so those are the kinds of use cases, but in all these use cases, we definitely have human in the loop, namely human decisioning. None of the use cases so far are just like agents go do your thing and, you know, we'll hope for the best. It's it's definitely human decisioning at various checkpoints. >> Yeah. And and so I mean I have an agent, you know, following my emails, uh, and it it writes drafts and then I go through and and approve and send them. Is it that sort of thing when you say human human in loop that the agent in, for example, a loan application would do all of the necessary due diligence and and tee it up for a loan officer to to look at. So all the information there and he just makes >> one the one difference between what I'd say uh in a in a consumer personal productivity where you can have an agent look at an email summarize it um and you know you can review it before you send out let's say email to a prospect a customer to a friend whoever it might be in the world of enterprise uh looking at this you know for example a loan approval you have to use look at the consumer you have to look at unstructured content like you know their statement bank statements, other information looking at the financial profile and then look at your company data database essentially your knowledge database to say based on the profile and the risk profile what is the APR that you might you know loan uh loan rate you might provide so you'll have to look at different systems systems of record as well as knowledge and then may kind of provide some uh um suggestions and that's where the human comes into play where you uh earlier a human would have to do all that due diligence figuring out all the details but the one caveat is here then you're taking action once the human approves the agent is then taking action to to process it as well. So it's not only about the knowledge angle of it to get information but it's also the action piece of it and that's why there's a higher level of governance and guardrails because ultimately then you're processing it through these financial systems or healthcare systems from a transaction perspective and that's where you know agent tooling for example comes into play where you can you know run a process in SAP or run a process in whatever other system and agents have that capability. >> Yeah. Do you do you see a big difference uh because I'm hearing this from people uh between uh companies that have legacy uh deterministic software running uh their automations uh and and uh AI native or agentic native uh startups that are building everything from scratch. I'm sure you're aware of Sam. >> Yeah. bet on the first uh single person with an army of agents to build a billion dollar company. I mean that's >> Yeah. >> Yeah. I mean so what's exciting about the category we're in is you know you have the big enterprise vendors who are getting into this broader agentic automation category. You have the new startups AI native startups. Um and it's not either or. I think depending on the use case we believe at automation anywhere it has to be a combination because when the AI native startups you know I'll give you one example a uh when a let's say um start AI agent you know AI native startup like NA10 whoever will talk about concepts or crew.ai AI let's take AI we'll talk about concepts like agentic orchestration but what they mean by that is that the agent is is deciding based on intent how to orchestrate and so it's all you know cognitive it's it's probabilistic right >> and whether it's agent calling another agent so remember one probabilistic step calling another probabistic step is going to be a lower uh you know potentially instead of 100% accuracy it might be even lower but you're literally leaving it to the agent to go figure it out and you there you potentially use cases for that and in some ways open claw is an example of uh doing that and that might work in certain scenarios in our world. I think one is the level of trust but also the other is frankly customers still you know have obviously legacy systems behind a firewall systems you know uh that where the deterministic processes play really well but also I don't think the agent technology is at a point where you can get 99 or you know 90% plus accuracy all the time and so there is a trust factor and you know we're still early days this technology is quickly evolving you know 18 months ago there was no consider agents. Now there's agents and the technology is fast um uh evolving. So today we think for these mission critical processes and I want to be clear not for every use case but for these mission critical processes that drive the these business outcomes around operational productivity or regulatory compliance or new or revenue impact. I I do think we believe it's a it's like combination of deterministic and cognitive. Today we feel it's the 8020 rule. That's what we see with our customers. Will it become more 50/50 6040 and eventually 2080? Time will tell. I think it will evolve. It'll definitely evolve. The speed of that evolution though uh it's not only about one is the technology. Second is the governance and compliance around that technology but that's also evolving things like you know agent of valves still evolving. I mean in terms of figuring out how will that agent actually perform and then in production does it perform what you thought it will perform because those things are really important in you know especially in a you know business enterprise environment and then third is the just human trust like do I trust this uh system this cognitive system just to do everything without any human intervention that's where I think it'll take a little bit time in in you know industries like healthcare and financial services and other in kind of those kinds of industries but I think it's starting. >> Yeah. Yeah. I wanted to ask you about orchestration process intelligence >> and predictive resource execution how those enable uh mission critical automation can and how automation anywhere uh applies those things. >> Yeah. So we the core um um kind of central nervous system is the orchestration because that is what orchestrates across these legacy systems the decisioning within that does it need to call an agent does it call an API a you know a bot you know process some documents whatever it might be uh and I think the benefit of the having the orchestration engine our engine is called the modzart orchestration orchestrator essentially it it can automate across any application in any system as connectors pre-built connectors customers can create their own but also you have tools like MCP both inbound and outbound for you know to call a process to be orchestrated or to call another tool which may be another system you know through MCP outbound so so it's kind of the core system and which where you can have a mix of deterministic steps agents and human intervention all that every time a process runs so that's core engine, right? So every time a process runs though, there's a lot of data and metadata captured. So we have over 400 million processes at any given time running through our system. So while we at automation obviously don't look at the business data flowing because that's a little bit irrelevant for us, we can see and it's also by the regulatory compliance thing that we can't look at the business data. We can see the metadata flowing. We can tell how an audit to cache process works. You can tell how prior authorization process works. You can tell how AML process works. So it's using that we've created a system called the process reasoning engine or with the goal that how do we improve the outcomes for a process. So when our agents are performing not only do they have you know the tools and the goals and the memory management that they need the process reasoning engine so that past behavior of those agents can be used to improve the outcome. the next time the agent performs and that's kind of our focuses which is not and so it's not about general purpose agents it's really about how do we enable these process agents as part of these you know longunning processes powered by a Mozart orchestrator to have you know higher efficacy outcomes and lower the hallucination. That's where talking about you know how do we use the metadata or the data that's really the process reaching engine which is driving that and then obviously we have the full governance and compliance and things like age intervals um masking of PI data so the full governance piece is also really important because the first question we get from customers when they're looking at any kind of agentic processes what's the governance and compliance because we got to make sure not only the CISO but the legal team are fully aware because often customer data is flowing through these processes. >> Yeah. And on the Mozart orchestrator, is that uh deterministic or or is that uh based on uh LLMs and and generative reasoning models. >> So they're the agents are based on you know they have gold we have goldbased agents and reasoning all that stuff. uh but the what's that orchestrate itself is a deterministic process that includes the ability to embed agents as as a as an action. So it's basically think of agents as a first class citizen within the Mozart orchestrator. Most of our customers will build when they do an order to cache process a prior authorization process they build a process as part of Mozart orchestrator. Now it doesn't prevent a customer from using an agent directly as well but we see less of that because they want that control of a kind of a deterministic uh like I said a central nervous system where they can control how the agents are being uh used and they use the agents at the right time. That's the main thing. It's where the agents are used at the right time. You're not using agents for everything. So for example, if you want to, you know, uh get some information from SAP and you know you have to get that information, you don't need to use an agent for that. You can use an API for that. >> Yeah. What is predictive resource execution? >> So predictive no so it's not it's called process reasoning engine. That's what it is called. But essentially it's a a it is using conceptually we have a lot lot of metadata and using that metadata we basically at the core created fine-tuned models for process use cases. So example we have fine-tuned models for how do you extract and process a contract. We have things for order management. we have different you know uh legacy systems that our customers interact with. Uh so how do you um automate those legacy UI systems to through traditional RPA as well as how do you build processes? So there are lots of different models that we have created on top using that metadata and then there's an entire context graph around it which is saying okay now we've created these generic models built on LLMs but it has to be contextual within a certain within a customer's specific business. So for example a though you know a manufacturing customer is going to be really very different from a financial services customer. So that's where the context graph comes into play which is a combination of their business data could be their SOS systems or records could be their knowledge graphs but also all the processes that are running and the agents that are running that is also context for us and so it's combination of that and then you add you know memory management and um um and um self-reflection basically learn constant learning that is what we call the process reasoning engine and again the goal there is how do you improve the outcomes of these process AI agents because uh you can build it but it has to continuously evolve and because these are highly probabilistic in nature. >> Yeah. Okay. And yeah I'm sorry I had P I had the wrong uh >> that's fine. No worries. >> Yeah. Yeah. Uh so um how is AI transforming traditional uh robotic process automation into more adaptive and autonomous systems? I mean you've talked about that already but u where does automation anywhere see the biggest impact? I think the the at a at a macro level AI agentic AI and we you know we believe this that's why we've gone all in on this it's completely transforming the industry and our customers in a couple of different ways one at a core we believe that customers can now automate more the reality is with traditional automation whether it's RPA or just kind of you know process automation there were limitations because you could only automate what you knew It was very well you have to be it had to be very well defined. Uh but now with when you combine what you know with what you may not know with the exception uh scenarios now you can use agentic AI to deal with those exception scenarios or deal with scenarios where let's say you have unstructured content. And so that combination at the core allows us allows our customers to automate more of their you know operations which is I think a big deal for them because ultimately leads to you know real business outcomes and I think from a technology standpoint I'd say there are big there have been big benefits of our customers. One example I'd give you one of the challenges with traditional let's say RP or robotic process automation because essentially it was mimicking user behavior and so it was going into let's say logging into an SAP and automating that screen and and blah blah blah but if that screen changed then our the bot would typically fail right because ultimately the screen's changed the experience has changed but it the for the field is no longer there or new fields are there we rolled out this product called generative recorder which really uses vision models as well as generative AI but also vision models and we have a we have a lot of understanding the DOM structure because we have this history of RPA so we created a model around that a fine-tuned model now we see we enabling our customers to drive 60% higher resiliency of even pure play RPA that's just one example there many other examples like building processes using natural language but essentially allows uh you know higher resiliency uh faster time to build, right? You can build these processes and you don't need to be a expert to build these processes. You can build it with natural language because you know what you want to build but you don't even need to know the technology behind it. So the um the barriers to drive business change have lowered tremendously and you can see this you know whether it's whether it's u open claw or other or you know um uh cloud co-work which is in preview beta all these tools are reducing the barriers as far as you and I can just you know anybody can go build these and that's the biggest benefit I'd say for our customers because frankly you don't want a central IT team to be the only ones to do it. You want the line of business and their folks to be able to drive efficiencies and outcomes within their own business. That that I think is another big benefit. >> Yeah. Um can you talk about how how does you how do you work with a customer? I mean say that uh you know a highly regulated uh company in a highly regulated industry say a bank or a hospital uh comes to you uh and they have you know some low-level automations in place uh but they they want to uh expand that automation with a gentic workflow. So how how do you what do you go in and do an audit initially or or how do you work? >> So one is we've um the big three industries that we work we've traditionally worked with our financial services, healthcare, manufacturing. So we have a lot of data, a lot of history uh and forget the technology used but we have a lot of history with the outcomes that we can drive. Uh so typically we have this process called the BVP or business value proposition which really looks at outside in looking at maybe the financial statements and other area other areas where what are what do we perceive as the potential areas to impact areas of impact. So what are the outcomes we can drive? But that's an outside in view right because we don't at that point we've necessary not necessarily got detail in depth with the folks within the company uh to know more detail what rack actually they they care about. So you have outside in view and then we work with um the the key influencers whether it is somebody in the CFO's operations department the CFO somebody in line of business to really then map that to what are the outcomes that they care about because all of us have outcomes that we care about and some other outcomes are not as high in priority because we always want to look at that first couple of areas or use cases to say this use case actually there might be benefit because it has a mix of okay there's a process maybe a longunning process but it has unstructured content and it impacts maybe cash flow or there's some outcome the mistake I think folks made at the start um of this entire push into generative AI and agentic AI was it was folks are trying to use it as more as a technology tool versus thinking about okay what outcome we want to drive because you know the technology is cool if it doesn't drive an outcome at least for our customers you uh it's kind of useless at that point because what am I going to do in an enterprise setting? So that's a first step a BVP with trying to identify the top two or three outcomes. We always have a you know buyer or or a influencer who kind of owns their outcomes in the customer side and then know we obviously work with it. We do know PC a solutioning uh we have hands-on sessions for the technical teams where they can actually build using our platform. So it's kind of a multi- uh level effort and that's typically what drives you know the selling motion and then post sales what we found what now we're uh especially last 18 months what we realized is AI and agentic AI is not the same as deterministic right deterministic we could allow customers just go do it on their own frankly or maybe a gsi helps them we're very hightouch with agentic AI because even you know defining a goal and define the right tools and kind of optimizing that output. It's not easy for our customers. So we have this concept of pilot to production where we help our customers roll out the first set of use cases and train them while we're doing it. And so while once the first set is successful then they then they can take it forward. So it's the end to end. It's not like a traditional what I was you know used to at Salesforce where you're selling the software and then yeah you have services but but you're not thinking about it end to end. Now I think all of us have to think about it end to end because a customer is not just going to be able to go figure out how to deploy these agent requests. >> Yeah. And that's something I see a lot of these agentic uh builder platforms that are being pitched to enterprise. Uh unless you have someone that really understands uh a gentic AI on the customer side uh just having the platform without some guidance is uh doesn't doesn't really solve the problem. And and as you said when I mentioned using OpenClaw uh you know there are all kinds of security vulnerabilities or potential security vulnerabilities and and you want a company that understands that uh to to make sure you're you're not opening your yourself up to to things. So what are the challenges that enterprises face uh when deploying agent AI at scale? What do you see the main challenges as? >> I think the couple of things. So one we just may talked about I think the this technology is evolving so fast. >> Yeah. and so much hype sometimes and there is also a lot of vendors who who interchangeably use the term agentic AI assistance chat bots like it's like all like confusing >> and but everybody's also talking about agentic AI from a different slightly different vantage point there's a different lens right how Salesforce talks about agent force is really focused on salescloud service cloud their applications how we talk about agentic process automation is from a different lens is a process lens a service now we'll talk about is a little bit different. So everybody's a little bit different how they are approaching uh agent agentic AI. So that's one uh that's one big challenge to for our for a CIO to understand okay what do I use because you also don't want like 100 tools that are all having you know silos of uh application. That's one. But once you figure that out, and I think that's the the forwardleading CIOS understand when to use what. The second is, you know, I think things like pilot to production are really important uh to enable that deployment. Figuring out the right outcomes that you want to go after versus treating as a technology problem is super important. That is like a make or break for companies and those who know what outcomes they want to go after. And there's a line of business and technology at the table. similar problems from the from the past is no different from when we were when we were talking about just deterministic automation because even deterministic automation of processes or or deals if you don't have a problem you're trying to solve you're just using for the technology it's not going to work but it becomes even more now because you know there's a cost of tokens there's a talk cost of using LLMs and you know use um uh outcome there's a pressure from the top to say okay what what are the good outcomes that you're seeing Um and then I think if you go to the technology there are there's a definitely a knowledge gap for some customers. For example, what type of data uh context do you want to use for an agent? >> Yeah. >> Ask what we'll see is okay here's all my catalog, my SharePoint servers, my knowledge. Just let's put it all into a rag and it'll all be good. Well, that's not how it works, right? Because if I'm doing an order management process, I don't need to know what's happening in, you know, uh in another area, right? But uh I I might need some product catalog information. I may need some order information. I need may need some, you know, shipment information. So I the kind of context that you or or systems of record knowledge articles that you use is also very important. So you need to build the right context for different agents or different workflows uh versus trying to put it all into one big rag and you know see how how it goes because sometimes more data is actually worse for it because it actually know the efficacy goes down south. So those kinds of um uh that kind of understanding I don't think everybody has that and how and that's why we have these pilot to production where we go deep in pretty detail with the customer to say how do we actually build this and what is the right context to use when you're building these agentic processes right and then that's a critical element of success uh for these and then obviously the right governance. So how do you put the right governance in place, the audit in place, make sure the compliance team is gives a thumbs up because they'll want to get all the audit logs uh in terms of what what customer data went into the agent, what how what is the how do the agent you know what is the plan, what's the reasoning, how do they actually execute and the response. entire end to end has to be well defined because you can easily get into uh compliance issues and legal issues if it's not if you don't track some of those things. So it's really looking at the end to end around it. So those are some of I would it's a challenge but it's like something that is quickly evolving. I think as the the the key thing is as customers deploy that first set of use cases. This is where some of these challenges get ironed out because once you set the groundwork then it's easy to scale out. Yeah. Um I wanted to ask um you know I I talked to consulting companies the the big uh consulting companies and there was a day when people like Automation Anywhere the the RPA providers uh were were what the consulting companies would bring in to solve uh problems in in whatever business case they we're looking at but increasingly consulting companies are getting the business of building solutions and it seems like you guys are getting the business of consulting. Is there is there kind of a merger between those two worlds do you think? >> I mean this like I said uh my the all these industries my industry and my is all getting disrupted. I mean we're going this concept of software as a service to service as a software. I I don't know I mean different folks have used different terms but >> yeah I think but we're also partnering with them. It's not that we're doing going it alone. uh we partner with the biggest gsis because sometimes we're like we can't we're not a services company right that's not our we're a product company we're a software company >> right >> so while we will have enterprise architects and those who are you know all in all the details of how to build these AI systems we'll partner with the gsis because they can you know provide those services at scale so it is a collaborative exercise it's not a eitheror but this industry is changing right even think about BPOS's how BPOS used to operate is about, you know, people and time and money. Well, with AI, do you need as many people to do even even more deals that you might get? No. I think they're heavily using a lot of our customers are BPOS or MSPs and they're using our agentic processes to drive higher operational efficiencies for their customers that they manage. >> So, all our industries are getting very um you know disrupted. But I think it's a collaborative exercise with these GS. We don't want to be in the full-blown services business. That's not something that's >> Yeah. And and one of the advantages I can see of of working with Automation Anywhere, as you said, this stuff is is, you know, and I'm focused on it, and even my wife will mention something that I've been hearing about, but I haven't taken the time to read about. I mean, there's just so much new stuff coming at everybody. Uh I would imagine you have a team at least that spends a lot of time tracking u advances in in a gentic or generative or whatever >> part of the AI stack it might be. >> Yes. With the with the caveat that all of us have to do it. It can't just be a team. Um >> I think the old days where you would have a strategy team they would look at okay what are competitors doing what's the industry doing where should we going maybe some big companies might still have that super large but I think those it's moving so fast I I can't wait for a strategy team to go figure it out all of us as an R&D team all of us are trying the latest tools we have very strong partnerships not only the hyperscalers but also with openi anthropic and certain other startups who are um who are at the leading edge of these models and other other pieces. So we're actually actively using some of these products even before they go generally available even from those vendors. So we get to see like operator AI when it came out from open AI we played with it because that essentially is computer use technology right which is essentially could uh enhance RPA at its core we could we were able to use you know open AI anthropic we have other startups we work with so we're always trying to be at the leading edge at least six to nine months ahead before it actually the general market is aware of it uh because we also have to see what technology applies Because one of the key things is there's going to be a lot of noise and lot of new innovation but not everything applies in our world. Something some may only apply in in a certain industry or in a consumer may not be applicable for our use cases. So it's really also uh important to understand what applies and where is it in its technology evolution. Is the time to incorporate it now or should we wait a little bit till it matures a bit till we incorporate because we have a higher barrier so to speak with our customers because it is those mission critical processes. Yeah. >> What's good for for consumer personal productivity may not be good for those mission critical. >> Yeah. uh and you guys recently announced an integration with OpenAI. Uh can you talk about that? >> Yeah, so uh from the day one like we've been working with OpenAI 2022. Uh in fact our a lot of underlying LLMs use you know um open AI as models no sometimes through open um others through Azure we also use Gemini we use anthropic cloud anthropic is a big partner of ours uh and obviously AWS is you know we are SCA um partner basically their top 1% partner so very strong partnerships because you can't do this on your own but around open AI we started building these joint solutions because essentially We have the again the process context. OpenI has best-in-class models and we're saying for example for prior authorization or there's you know there's another one around you know um in financial services there's a use case how do you create end toend solutions but using open AI as the underlying NLM for those agents to act. So those are the joint solutions we've been creating and we announced some of those I think um a month ago and you'll you'll hear more of it as we get to our imagine event in May. >> Okay. Um the looking forward I mean at this point uh I mean you talk about missionritical automations uh what percent of your most aggressive customers what percent of their business processes are uh are automated using Aentic AI at this point I mean if or or whatever metric you can use And how do you see that growing? And there's a lot of talk about you know uh you know an automated organization I mean a a a gentic organization where you have basically uh you know strategy at the seuite strategy at the top maybe working with an AI uh reasoning agent to to direct strategy and then below you have human managers that have teams of agents. I mean, do you see that happening anytime soon? And and how far are we from that? >> I think so. Our term for that is the autonomous enterprise where, you know, um uh so I think that's maybe that that's a goal that we suggest to customers now three-year journey. It's again not about replacing humans there. It's about putting humans at the center of decisions but where the operations are running through again in our world through deter combination deterministic and agentic where you don't need to you know for humans to do the the uh minutiae in terms of different tasks right they actually making decisions at the right point uh and where we go from just one department to pretty much multi- department across departments in the entire organization because right now we're still department by department. That's where we are. And even if you say um you know as I mentioned 1500 odd agentic processes and we have many more in uh in in pipeline where customers in pilot looking to go production. It's still the first set of departments. It's not like they've gone wall to wall with it. It's still in that phase where we're trying to see they are the first set of successful use cases. Now they want to expand that more and then they'll go to multi- department. So that's the phase we're in. So are we so the point where we go into this full autonomous enterprise I think is still 3 to 5 years out in my opinion and it's not only about the technology again here it's not only about technology it's also about the appetite for risk the uh obviously the evolution of the technology. The technology has to get better. The governance has to get better. the things that need to get better which are getting better. Um and then also change uh a lot of this is change management within an organization that's not easy and so a lot of the customers who've done this are early adopters you know are innovative so generally the head of transformation is a great customer for us because they're looking to transform their organization right so it's a great buyer and somebody who's willing to push because um there can be inertia with these within these large enterprises um but to your answer you your question I think we are it's a three to five year journey for the full autonomous enterprise I think we have lots of customers as I mentioned who are who deploy their first set of use cases but the goal over the next couple of years is to expand that to not only within a department but multi-ep department and you know uh areas of the business maybe uh in some in one customer they're looking at just within uh uh UKI and then they want to expand it to all of Europe so there are various ways that customers will look to expand these agentic processes um you know depending on their setting >> change management but one of the issues in that is that you've got employees that have no idea what what AI is or how to work with it. Uh there's a consulting company that I talked to and they're developing something called the the uh agentic quotient, you know, sort of like IQ, like identifying or being able to measure employees uh and identify employees who have that aptitude for working with agents and you're that's you're going to need that. So how do you manage that or how do you talk to enterprises about that? I mean even our own organizations we have to do we've had to do that right just because we are a software company in the heart of Silicon Valley it's not that we woke up and everybody was like let's do AI I mean more so the product engine we've all always you know into it um but the rest of the organization we we have a finance department we have you know different departments that are going to be you know they need a little bit more prompting so to speak to start uh getting there I will say a couple things one is I think one thing chat GPD taught us is the barriers to actually use. You don't need to know prompt engineering. I I don't agree with folks when they say that hey you need to know how to prompt and yeah but I think if you look at whether strategy earlier now you look at cloud core work the barriers to do some of work with AI is you know going it's really decreasing day by day I would say and it's going to only become easier. So if you are somebody a nurse practitioner, I don't think you need to know how to prompt uh an assistant anymore and I think you can get work done. I think what what it will be even a year from now is going to be very different from where it's today which is very different from what it was a year ago. Right? So that's one I think the technology is also and the experiences are also evolving. I think the second though is um we do do there at least vendors like uh myself u uh and I'm sure other companies also do it. I think hands-on is really important versus just theory. >> So one of the things we do is as part we have a pathfinder community which is our basic our customer and partner community. We regularly have hands-on workshops where they get their and walk them through actually building an agent, building an agentic process, how to think about security, governance, all those things we just talked about and getting hands-on not only for the technical teams but also for the IT teams who are maybe not super technical, right? So that so they can actually you don't need to be a technologist or you don't need to be a developer to build and so that mindset only shifts once you get hands- on with it. Uh and so that's another way that at least we and other vendors have been trying to push um our customers and you know recently we had a pathfinder summit uh and it was amazing. We had over 40,000 um you know it was a virtual summit who attended and we have over 100,000 who completed a certification some who attended and some after the fact because is virtual we put it on our on our community site they completed after the fact and that's what drives more you know knowledge diffusion uh and that's I think something that combined with I think the barriers changing with uh the barriers to use changing with you know the co-pilots and the clot uh co-work type of examples I think invariably folks will start getting more comfortable uh but it takes effort it's not that I think the first reaction is always oh this is hard but as they start using it I think it'll get better >> yeah yeah uh yeah well that's fascinating so you have this program for for customers and partners uh to to develop It's critical. I think if just if you just think about it in terms of just software and they'll figure it out, I I think that's a miss because it's especially for our customers who are used to a certain way of doing work and building stuff. They don't get hands-on because there everybody talks about in in the same way, but until you actually use it, you don't realize that the context is different. >> Yeah. >> Also build the basics of you know using and building these and monitoring these. Again this is happening so quickly. Uh it it's also transforming as you said uh automation anywhere you know you also have departments that that are being transformed. >> Uh is in that it's moving so quickly. Uh does your relationship with customers continue? I mean it's not you set them off and set them up and send them off on their way. I would imagine because the underlying tech is changing and your approach to it is changing. So do you have kind of ongoing relationships with uh with enterprises? >> Oh 100%. So we have various a we have a community. So our customers are very open and giving us real-time feedback and how things are working. We have things like product clubs, user groups, we have a customer advisory board for large customers. uh we have our imagine event which is basically our major event which happens in the US and then other regions. So lots and lots of customer touch points including you know uh oneto ons through customer advisory customer meetings. So there's a lot of input and then we have real-time adoption data. One of the benefits of automation anywhere like my PMs will look at adoption data for a product from the day it's launched for the first set of customers who are the early adopters to the next set and we look at the history of that saying is it trending in the right direction. We'll do interviews. We'll figure out okay what can be improved. Uh and so there's an active handshake going on out there because this is fast evolving. So it's not that you build a product, put it out there and then wait for some time. We're constantly evolving the offering and that's happening because customers are giving us real-time feedback both directly but also through the adoption data. >> Yeah. Uh it's uh it's fascinating also a little intimidating uh I mean from your point of view is it uh is I mean I'm sure it's exciting what's happening but is is it also sort of uh intimidating because how do you get your arms around this thing that's that's moving so quickly? I think initially in the early days it was a little bit intimidating because part of it is it's moving so fast there's so much new technology and then some of stuff is not real like at some point there's this um term around large action models >> somebody and essentially now that's a in some ways agents are action models or you know that's what they do >> but that wasn't real. So you also have to decipher what's real versus not real and that you can only do by trying it by actually using it. Um so there was a little bit of that but now frankly it's all about excitement. It's as a technologist that's what I'm at at my core and my team they are super excited because this is allow this is like um rapidly changing market rapidly changing um uh technology but also the opportunity to if you use it in the right way to really drive customer outcomes that we could not have driven before. So we're defining the future in many ways. I mean defining the current and the future very rarely do we get that opportunity as technologists to do it. I think it's the uh after a long time I think the last being maybe the shift to SAS um maybe the cloud or even before that the shift to the internet um I was I was too young to remember that but I think this is the m big moment for us and I think there's more than um I think excitement is the right kind of mindset for us. >> Yeah. And do you think that uh companies I mean you you talked about early adopters uh are companies if they wait too long are they going to be uh playing catch-up? Are they going to be facing uh competition that that is uh stronger than they can handle? I mean as with the internet the transition to the internet I remember a lot of companies were slow to to adopt. Yeah, I think it'll obviously depend on industry because there are certain highly regulated industries that like defense for example u take a little bit more time but broadly my broad statement would be yes I think there's also a learning curve on what good looks like so if we had not as automation anywhere not started four years ago on this journey we would not be here if you just started a year ago and woken up one day and said oh now let's get on the agentic bandwagon and let's go make it it wouldn't be the same because there were a lot of learnings as well as missteps in term as we went through those learnings that helped us be what we are today. And that's similar with the customer where if they start even they start small where they are day one or year one versus where they are year three is going to be very different, >> right? And so they have to get on that journey but obviously with the right you know framework and the organization structure and the compliance framework because trust especially in these large organizations is super important. AI is moving fast within the enterprise. Employees are experimenting with personal AI accounts. Teams are building custom AI apps and autonomous agents are connecting to sensitive systems. Innovation is exploding, but governance isn't keeping up. Gartner predicts that through 2026, 80% of unauthorized AI transactions will come from internal policy violations like oversharing sensitive data, not external hacks. That's a huge risk. And that's why Island, the creator of the enterprise browser, launched its new AI services. island surrounds generative AI, AI browsers and autonomous agents with the enterprise controls that they were never built for. Identity enforcement, data protection, auditability, and centralized policy. Now, organizations can safely use any AI app, consumer or enterprise, without corporate data loss. Teams can deploy secure permissioned agents with full audit trails. Island's publishing capability makes it easy to safely share internal AI apps across the company. And the Island AI browser brings governed multimodal AI directly into the enterprise browser employees already use. The result, a single secure foundation for AI at work. So CIOS and CISOs don't have to choose between productivity and protection. To learn how to scale AI without losing control, visit island.io. I O.
📺 Ähnliche Videos
🔮 MiroFish simuliert die Zukunft mit tausenden Agents | Alles was du wissen musst!
Christoph Magnussen · 2026-05-11🇩🇪 DE
Ich habe das fortschrittlichste KI-Tool der aktuellen Zeit entdeckt… Das ist Hermes Agent 🔥
Der KI-Doktor · 2026-05-09🇩🇪 DE
Diese OpenClaw MasterClass Wird Deine Arbeitsweise Für Immer Verändern
Der KI-Doktor · 2026-05-08🇩🇪 DE
Paperclip Is Insane | Full Tutorial
Ferdy․com | Ferdy Korpershoek · 2026-05-06🇬🇧 EN