Name: Is AI Rewriting the Rules of VMS? – The Future of Extended Workforce Platforms
Uploaded: 2026-07-28T18:31:20.844Z
Duration: 1 h
Description: Is AI Rewriting the Rules of VMS? – The Future of Extended Workforce Platforms

Transcript for "Is AI Rewriting the Rules of VMS? – The Future of Extended Workforce Platforms": Hello, everyone, and welcome to today's session. Is AI rewriting the rules of VMX? The future of extended workforce platforms. So I see people are starting to join and and get set up. So while everyone is getting settled, I'm gonna run through some of our standard housekeeping items. So for those who want to watch this back or need to hop off early, we will be sending out a recording of this session within twenty four hours. So keep an eye out for that, in your inbox. We would love you to submit any questions you have, so please enter those via the q and a tab on the on the side of your screen there as we go through. We'll try and answer as many as we can live, during the q and a at the end. And, of course, we're always looking at ways to improve these types of sessions, so we'd love for you to fill out the feedback survey at the end of the webinar today. Then please take note of, our product statement on screen here. This applies to any forward looking features and road map items we'll discuss today. So with that, we'll dive straight into it. Today, we're tackling a massive shift in the extended workforce space. The traditional vendor management systems are fundamentally engineered as static systems of record and linear workflow engines, But with the rapid rise of embedded intelligence and autonomous agents, we have to ask ourselves, is AI completely rewriting the rules of the VMS, and what does the future of the technology look like? My name is Catherine Flynn. I'm one of the senior product managers on the Venli team based over here in Ireland, and I'm delighted to be joined by two incredible experts to explore this topic at length. On the line with me today, we have Krishna, vice president at Everest Group, bringing fresh insights from Everest twenty twenty six VMS peak matrix research, and also John, our head of product management at Workday Venley. So welcome to you both. Great. Great to hear, Catherine. K. Thanks. Thanks, Catherine. Thanks for having me, and looking forward to the discussion. Great. Thanks. Always a pleasure, Krishna. So here's our agenda for today. We'll kick off with Krishna breaking down the macro market, shifts, then we'll look at the essential data foundation required before AI can even operate. And from there, we'll unveil how agentic orchestration is turning the VMS into a system of action, including a quick look at some real AI use cases, and we'll conclude then with how total talent management is changing in the age of AI. And so let's start with the analyst perspective. Krishna, you know, based on Everest Group's research, how are c suite executives rethinking workforce execution right now? Great. Thank you. Thank you, Catherine. And I think, perhaps is a good place to start the conversation. So to, to give our, listeners and attendees a bit of background. So we at Everest Group have done a series of surveys recently. Right? So we've done this with CXOs, talent acquisition HR leaders, procurement and contingent workforce category leaders. So we've done it across the board. So the objective really was to understand how the priorities of these different stakeholders with respect to talent and workforce, are changing, are evolving, especially in the current day and age of AI or more so agent AI. Right? So what we're witnessing across global enterprises, global organizations is a fundamental shift in executive priorities. So the C suite conversation has clearly evolved from a tactical recruitment question. So they're not really asking how do we fill roles. So the conversation is moving towards the orchestration question. Right? So which is essentially they're trying to understand how do we assess, allocate, and orchestrate work across multiple talent channels. And, of course, you know, operate operationalize this across talent channels and simultaneously as well. So a few findings, right, that probably all of the attendees and, the panel here would find interesting is so 55% of the CXOs that we have, we have surveyed rate moving from static workforce models to flexible blended talent models as critical for their workforce strategy going forward. And similarly, close to 40% of respondents expect channel diversification to be a top priority over the next two to three years. And a similar number of respondents, close to 39% respondents, in fact, explicitly called out digital labor, which is driven by specialized AI agents to handle core operational workloads alongside human talent. So I think from an analyst perspective, these numbers, these are really substantial for us. because if you really think about it, two in every five CXOs are prioritizing AI agents. And this is also happening at a time when we have been in the AI hype cycle for a better part of the last two to three years. Right? So clearly, this is not a knee jerk reaction, but a carefully calibrated response and acknowledgment to the changing market dynamics. Yeah. Yeah. And lastly, we also saw 78% respondents rate shifting from reactive execution to predictive data backed decision making as essential to their future operating model and for and for their future of work, for their future of work priorities as well. And you've also been focusing in then on the contingent workforce as part of this reach research as well, and we're continuing to see movement in terms of the importance of that work worker population. Excuse me. So could you share more on on that particular strategic shift? Yeah. Absolutely. Absolutely, Catherine. So in fact, you know, once again to reiterate, so some of these shifts that we're talking about, these are not point in time shifts. Right? So these instead point to a bigger transformation that is well underway. And clearly, CXOs are approaching talent and workforce strategies and decision making more holistically. And that's precisely the place where continued workforce management has clearly become more important across organizations. So clearly, contingent workforce has been elevated from a back office procurement task to a primary strategic growth lever for organizations. Again, some more statistics on this, on this slide. So 70% of contingent workforce program leaders that we surveyed tell us that contingent labor strategy is a primary driver for enterprise workforce agility. Right? So if I ever take think of an example here, if you really think about what happened immediately after COVID, especially with the medical device manufacturing industry. So clearly the demand for medical devices increased post COVID. Right? So there's a lot of demand for, testing kits and other, equipment, right, related to COVID care that has become that has been in demand because of which the hiding demand across a lot of medical device manufacturers has shot up. So I think that's a place or that's a time where they have shifted their focus towards continued workforce. And, clearly, it has really enabled them to flex up and down, right, as per the demand, they're witnessing in the market as well. So, clearly, continued workforce provides that level of, agility within your workforce to be able to flex up and down as needed. So So that's one thing we are seeing. And and, of course, you know, 58% of the respondents also said they rely on continuing workers for scarce, niche, and specialized skill sets as well. Right? Especially in places where or especially for skill sets where it becomes very difficult to hire through the permanent workforce channels. So what's important to note here is we also have to look at these statistics in conjunction with some of our earlier findings on channel diversification as well. So when we're talking about channel diversification, essentially as an organization, you need to be able to unlock access to all of these specialized skill sets. And of course, you have to be able to tap into various channels, right? Including temp labor or statement of work, services procurement, independent contractors, so on and so forth. And lastly, we also saw 34% of the CXOs that we surveyed, viewing contingent talent from trusted suppliers as a critical as critical to their core project delivery in the next two to three years. So I think the word trusted suppliers is important here. So essentially, suppliers that can essentially become strategic partners and provide access to these proven skill sets, and also deliver the outcomes that organizations expect and matter the most becomes critical in this day and age. So the bottom line, as we look at it is external labor or contingent labor continues to become central to business execution. And, consequently, the underlying BMS can no longer just be a passive database or cannot just act as a system of record alone, right, for managing invoices or for managing requisitions. So it must guide strategic workforce decisions across skills, suppliers, channels, centers. So, so let me ask this to John, right? So, John, do you broadly align with, with our point of view here, or is this is there something that surprises you here? No surprises. And definitely from our from our product standpoint, as we talk to customers, we're seeing a lot of that transition from just tracking skin spin to really just guiding the strategy. So being able to dynamically add skills in your organization quickly, respond to that change is is really critical in today's market. Absolutely. Absolutely. Of course, you definitely nailed it, John. So traditional workflow first VMS design for contingent workforce programs, clearly, in our point of view is officially reaching its limits. Right? And if you think about the reasons, there are several compounding pressures that are stretching these workflow based VMSS to the breaking point. We have outlined a few of those forces here on this page, right, where we are talking about things like greater workforce complexity. So today's contingent workforce programs, when we think about them, there is a need for orchestrating various types of contingent labor, which also brings with it a lot of challenges and considerations for organizations to think about. So these can include things like managing hybrid worker types, thinking about global compliance, thinking about IC classification, reclassification, thinking about statement of work, deliverables, and essentially all of them are operating in one ecosystem. Then of course there are also challenges related to fragmented data and ownership. Right? So addition critical data, which is usually scattered across multiple technologies, multiple systems like the HRIS, the VMS, the procurement engines, and of course, broader enterprise applications, that also becomes a challenge, right? For you to be able to really elevate the outcomes that you expect from a, from a, let's say future proof or future ready contingent workforce program, you cannot really operate with fragmented data and fragmented systems. Right? Similarly, if you think about the demand for speed, so hiring managers also now expect instant access to skills, which also puts immense pressure on HR procurement and MSPs to compress cycle times and compress them pretty significantly as well. And lastly, there's also an expectation of intelligence and more so embedded intelligence within technologies or within the workflow itself. So business users don't really want software that merely routes tasks. So they are expecting proactive and real time recommendations, which can help them to take actions right, almost instantaneously or immediately. So I'll I'll just end with one final thought here. Right? So the real risk in today's, operating environment for VMSS is not really the workflows. Right? So we're not saying that the workflows will stop working. So I think it is, really working well to adjust to the speed, the insight, and the opportunity, the demand that that we are seeing, and how VMSs can sort of cater to it. That becomes really important. How quickly they can adapt. So we are dealing with significantly heightened complexities, and, of course, VMSs clearly need to catch up. Exactly. And then honing in on, I suppose, the technology side of things and that intelligence layer. Specifically, we'd love to hear from the audience, via the poll now, which I think will will pop up for you all. But we want to know where your organization stands today in terms of the biggest barrier preventing you from utilizing AI in your contingent workforce program. So is it things like fragmented systems and siloed data, security, privacy, and compliance concerns? Maybe there's a lack of leadership, buy in, or budget, or maybe you're just unsure of what's the actual ROI of investing in AI and and where do you actually even start? So we'll take a few minutes to to let people cast their votes. Let me see. Can I find that? Oh, yeah. I can see it over here. Any early thoughts on what might be out in front? So, yeah, we can see it there. I mean, from from the audience perspective, security, privacy, and compliance concerns, which, you know, John, I know you've spoken to many customers about AI and and what we're doing in that space. So does that surprise you at all or or match with what you see when you're talking to program teams? Yeah. I'm not surprised at all. So it's completely understandable, security and compliance are taking the top spot on this poll today. So, when you introduce AI into a highly fragmented tech stack, your compliance risk just multiply because your AI is lacking your your single governed source of truth. So that's exactly why we believe that you cannot have a secure AI strategy without having a unified data strategy first. Yeah. That just moves. yeah. So kinda kinda pulling on that thread a little bit. So when you really, think about, these these various lenses, and where you see these challenges that are most pressing for your organization, a lot of them are gonna trace back to that same root cause. And and we're we're gonna double click on that a little bit next. Next. So, Krishna, enterprise buyers often ask us, can they just plug in an AI agent over all their legacy tech stack? What does the industry research say about that shortcut? Yes. So, John, I think the research in this case is unanimous. Right? So clearly, there are no shortcuts. Shortcuts don't work. So I think for a to work, you'd really need to have the foundational elements in place. So without a consolidated clean data layer across both full time and non employee workforce or non employee populations, a models deliver poor recommendation, which in fact also active detriment. Right? So it can heighten compliance risk as well. And, of course, we have discussed work based AI vision and strategy several times in the past. Right, John? So I've always found this approach to be intriguing. So if you can elaborate a bit more on how you think about it and how you approach it, that'll also be really helpful for our for our audience. Sure. Yeah. It it worked out. Our core philosophy is that the future is not gonna be constrained by what human teams can execute manually. Entirely new work workforce dynamics are gonna become possible, when every business unit is paired with a specialized AI teammate. But here's the catch. That AI teammate is only gonna be as useful, as the access it possesses in your full organizational context. And what I mean by that, it needs to understand your organization hierarchies, your finance codes, your security profiles, worker history, global compliance rules, and all that needs to be in real time. And so it really brings us into the next slide, and you probably hear me talk about data a lot today, but it really all starts with the data. That data needs to be AI ready for agents to be successful. An agent needs to not only understand what a thing is, but more importantly, it needs to understand what that thing is in the context of your business. So if you're feeding an agent just sort of the sort of rate cards, missing tenure dates, disconnected invoicing logs, It's simply gonna automate a bad decision a lot faster. It's gonna automate that bad decision at scale. Chris, does that reconcile with what you're hearing as well? Yeah. Absolutely. Absolutely. In fact, that's a critical point as well. Right? So ultimately, your AI is as good as the data you train it on. So in fact, in our peak metrics evaluations, when we talk to enterprises, when we talk to buyers, consistently, we keep hearing that data interoperability and unified governance to be the number one prerequisites before approving any large scale AI deployment in excellent workforce management. Yeah. That's precisely, I'd say, a different into the architectural approach that we've taken. We haven't just bolted the VMS onto an ERP, utilizing fragile third party APIs. So by integrating, unifying your Workday HCM, which is your employee system record, natively with Workday Vendly, who is your, extended workforce system record, you're really creating that single total talent data core. And so whether it's recruitment, your time tracking, learning, talent pooling, statement of work, that unified foundation is is truly what's enabling that total workforce superintelligence, and and really helping keep your compliance, exposure to them. Perfect. Thanks, John. So I suppose now we want to turn over to the core theme of the the session today, which is agentic orchestration. And, Krishna, I'd love to hear how is Everest Group defining the shift from basic automation to true agentic capabilities in the VMS space. Absolutely. Absolutely, Catherine. I think this is a very important topic. Right? Because in this kind of day and age, everyone claims to have agent decay capabilities. So when we really think about it, it all comes down to closing the gap between insight and action. Right? So essentially making actions more autonomous, more real time, and insights driven is what really determines whether you are going along the path of agentic orchestration or not. So this is a key ingredient that we look at, right? When we determine if we are really moving closer to our vision of agentic orchestration. So when we have conversations with enterprises, this is typically what we advise them on. Right? So the reason I say this is because if organizations want to determine if something is if some platform or technology is truly employing agentic AI or truly embedding it, This question almost serves as a litmus test while evaluating them. Right? So if you look at our market research, so it highlights a dramatic contrast. So interestingly, 63% of enterprise leaders that we have surveyed actually expect agent AI to have a transformative impact on their contingent workforce programs. Yet only 9% organizations report achieving true end to end workforce automation. Right? Why? Because AI can surface a brilliant recommendation. But if your underlying platform can't automatically trigger the workflow, if it can't sequence tasks across systems, escalate high risk exceptions to human owners and learn from previous outcomes, then the insight is not really translating necessarily into action. Right? And, of course, I want to also ask John for his thoughts here. So, John, does it broadly align with how we are thinking about agent orchestration and defining it? Yeah. Definitely. Your your AI can't be truly agentic if every step requires a human sitting at the keyboard typing in manual handoffs. Right? So the platform itself really needs to be ready to to execute in this new world. Absolutely. Absolutely. And, of course, you know, when we think about this, it also becomes important to think about how we can close that readiness gap. And clearly, it requires stronger data, tighter integration, and solid operating foundations. Right? So again, what is interesting to note is enterprises clearly have the ambition to scale AI enabled workforce orchestration. So if you look at some of these statistics, 89% of the CXOs cite digital automation or AI transformation as a critical enterprise mandate. However, at the same time, 84% of the respondents, right, contingent workforce leaders, also admit that their current contingent platforms, they remain disconnected from their broader broader IT ecosystem. So So I think this brings the emphasis back to the core elements that we were discussing earlier. Right? So having a stronger integration, tighter integration, stronger data foundation, and solid operating foundations becomes really critical. So if you're not really able to create this seamless interoperable ecosystem, it, agent orchestration cannot really be operationalized. Right? So you need to have systems talking to each other. You need to have systems across which data can seamlessly flow. Right? Again, we are talking about bidirectional flow of data and information across different systems. Right? So when we think about what this means for the VMSS of tomorrow, so the next frontier for VMSS is not just proving the basic operational value. So how can VMSS close this readiness gap so so that, you know, VMSS and these platforms can evolve from just, you know, workflow enforcement to dynamic workflow orchestration, that is really going to determine whether VMS will be an enabler of future ready contingent workforce programs going forward. Right? So I think this is an interesting slide here. So if you think about what the next frontier for VMSS is, when we're talking about this decision orchestration, we are looking at the various roles that a VMSS can play. And of course, you know, VMSS can fall into any of these distinct maturity stages. So as with any broad based system, so VMS today is acting as a system of record. So essentially, VMS is, you know, acting as that single system of truth, enabling, it's enabling basic transaction capturing, manual compliance, audits, and basic rules based alerts. So all of that can be enabled and operationalized by VMS of today. But what really brings about that differentiation VMS is going forward is if the VMS is able to scale up and get into some of these more and start playing some of these more advanced SaaS roles as we are defining them on this page. Right? So the next step, we have defined something we are calling as system of execution. So in this case, we are specifically referring to system of, let's say, a rule based execution. Right? So standardized routing, passing resumes, exception routing. So some of that, which is rule based, can be automated. So that's where VMS will start to get into the realms of being a system of execution. Right? And we also talked about the importance of data and analytics and intelligence. So VMS is also now becoming, more and more of a system of intelligence as well. Right? So we have co pilot guidance. We have predictive rate benchmarking and risk sensing, budget forecasting, so on and so forth. And I think we also did talk about intelligence at the point of decision making as well. And that's the place where the intelligence that VMSs can gather from multiple systems and help the users take appropriate action. In some cases, also autonomously. They start becoming or they start becoming, more and more of the system of intelligence products. Right? But what truly differentiates VMSS of tomorrow in our point of view is if the VMSS are able to truly act as systems of action. So, essentially, autonomous agent engines triggering a multi step life cycle, multi step life cycle of actions. For example, things like resolving policy bottlenecks, governing the decision execution at scale, so on and so forth. That is where VMSS are operationalizing or enabling through orchestration. Right? Again, a simple caveat to call out here, we are not necessarily saying that everything will be autonomous. We still value human in the law human in the loop. So humans will still be there to handle exceptions. And, of course, there'll be several judgment intensive tasks where the role of humans will be very, very important. But all in all, we'll see more autonomous execution of actions to the VMS itself going forward. I think I think system of action, probably, John, is also a term that's close to your heart. We have had some discussions in the past around this. So would really love to hear your thoughts on how you think about a system of action in the context of VMS. Yeah. Absolutely. That that system of action, and we'll go into this in a second, is is just a big part of where our focus is right now. And so let's let's break down a little bit of what what we mean by that, into the next slide. So as we're looking to operationalize this concept of a system of action, we're we're really focused on three complementary tiers. So first is embedded intelligence. So delivering smart contextual insights directly into the flow of work. The second is copilot mode. So equipping users with an AI teammate that understands, you know, in in natural language, how to help handle administrative multi step processes, but always keeping as as you were talking about that human in the loop to supervise and approve, as that works flowing through. And then finally, agentic orchestration. So really bringing to life that semi autonomous engine that offers a 24 by seven life cycle support, fully predicting, helping you plan, and orchestrate really complex life cycle steps. So we'll we'll drill into each of these, in the next few slides. So first, drilling into embedded intelligence. So with the data foundation, which we've talked a lot about here so far, we can start to embed AI powered intelligence directly into the flow of work. This means no more searching for answers. Instead, you're gonna have real time actionable insights that are embedded directly into your daily tasks, nudging you towards the next best move, without you ever having to leave your current workflow. In fact, we're working on this use case right now with our advanced vendor insights for job distribution. So today, a lot of vendor selection, is happening when posting a new job, the new job requisition coming out. It might be based on memory, spreadsheets, maybe around supports, a gut feel, a vendor who did well for you last time. So it's near impossible to use truly put your years of performance data to work for you, today and really have your vendors in every single job distribution decision. So plan to use AI to give your users guidance based upon all that historical data so they can make better, faster decisions to help them really improve that sourcing quality and the speed of filling those jobs. So advanced vendor insights will act as intelligent recommendation engine for vendor distribution. The system looks at how similar jobs and your vendor pools have performed those jobs in the past. We're gonna analyze things like your fill rates, your time to present, the applicant volume and quality. Based upon that, it's going to suggest vendors that are more likely to perform well for that specific job that you're creating. So, ultimately, the goal here is to help you route requisitions to vendors who are most likely to get you a quant a quality candidate that you need for that role and really cut down your time to fill. Yeah. I I think, John, so I think this is really a great example of how you can shift from hindsight reporting into moment of, into in the moment guidance, right, as we were discussing earlier. So, So, essentially, you prevent the operational mistake before it creates a lag. So showing rate or salary benchmarks across role types, and probably even locations at the time of requisition creation itself, this can lead to better outcomes in terms of cost control and perhaps even the quality of hires. Yeah. Definitely. And it really feeds nicely into the next area that we're, you you know, talking about, co pilot mode. And so this is where your AI is gonna drive, but your human is still supervising. And so think about you got those data insights up top. As you get into co pilot mode, AI is gonna take on the heavy lifting of those complex multi step workflows for you and really allow that user role to transform more into a person managing every step and the final decision maker. So if you can imagine engaging with a conversational assistant and explaining you'd like to schedule interviews with just the top three rated candidates on the job, and the system just really takes it from there and handles the rest. That's the type of simplicity that we're talking about here. So the goal of this was really to help eliminate a lot of that grueling administrative back and forth that consumes a lot of day of the the hiring team going through the process. So finally, this will all culminate in agentic orchestration. So this is where we really wanna move the DMS towards a 24 by seven life cycle support. Right? So it's, a semi autonomous engine that's just really always on. So the goal here is it's gonna be powered by specialized agents that are constantly working in the background. Their their purpose is gonna be to predict, to plan, and orchestrate, the end to end contingent labor life cycle for you from the initial requisition to offboarding and, as much as possible, minimal manual intervention. So it's really important as we think about AI governance here that we envision having configurable decision points where based upon organizational's policies, customers, can choose to force user to take action for workflow proceed or just capture an approval, you know, acknowledgement that the, agent should move forward. So, for example, imagine autonomous agent, constantly monitoring worker tenure, visa expirations. When a worker is 30 from their tenure limits, the agent may automatically draft an off boarding task or an extension and really notify the manager to come in and do that final that final approval. So the goal of this is really to help transform that VMS into a proactive autonomous system that's always on, really allowing humans in the loop to focus on the more strategic work. And, John, I think the that future use case of twenty four twenty four seven background monitoring, I think it's going to be a huge game changer for the streams. Right? Because most compliance breaches happen because a human missed a date, which is buried in the report. So autonomous agents can monitor and close this exposure or eliminate this exposure entirely. Yeah. That's the goal. So rather than just talking about it, let let's take it a look at one in action. So we're gonna share a quick two minute video demonstrating our contingent sourcing agent, which is in early adopter now, and it's gonna be delivered into general availability later this year. It's gonna show you how AI is practically grading candidates to help reduce your screening time, matching skills, and ultimately surfacing recommendations based upon your existing talent pools. With that, Catherine, you can start the video. The contingent sourcing agent brings powerful AI directly into Bimly, allowing you to grade applicants by matching skills to job requirements and instantly rediscovering qualified candidates within your existing talent pool. You are now looking at the applicant section of a job requisition in workdayvin.ly with grades displayed for each applicant submitted. This is the contingent applicant grading skill, which leverages AI to provide a grade in VINLY based on how well the applicant aligns with the requirements outlined in the job description. You can then navigate to the applicant's profile to learn more about their experience and skills. From here, you can then review the candidate in further detail to better understand the information that contributed to the candidate's grade. You are now looking at the second key skill of the contingent sourcing agent, talent rediscovery. This allows you to instantly locate high quality candidates your program has worked with previously, giving you visibility into their past performance and history so you can make faster, more informed sourcing decisions and reduce time to fill. Looking at the ineligible tab, these candidates are recommended by AI but may be unable to be applied to the job. This occurs due to information like rates, rehire eligibility, and candidate availability. From here, you can notify the vendor of your interest with a single click. This alerts the vendor to follow-up with the candidate, confirm their availability, and officially submit them for this role. Once you select send, a confirmation appears, letting you know that your request is on its way and the sourcing process is in motion. Perfect. So, Krishna, you know, as an analyst watching the contingent sourcing agent accelerate candidate screening and talent rediscovery, what's your immediate reaction? Yeah. I think I think what strikes me really is directly addresses one of the key priorities that we hear from enterprises. And, of course, when we hear this quite a bit as part of our peak metrics evaluations. So, essentially, the embedded explainable and outcome driven AI, is what enterprises expect because, like I said earlier, there are enough of experimental gimmicks out there in the market. So, clearly, I think this is something that, that that is of a lot of value for organizations. And so this level of agentic sourcing also will fundamentally shift the traditional MSP and BMS dynamic as well. So, essentially, elevating human teams from manual streamers to strategic decision makers. And, of course, we also want to acknowledge that this is not going to be easy. So, of course, we all want AI to, AI to operate as efficiently as it can and deliver the outcomes that we all need. So we do understand and we do acknowledge that this is not going to be easy, but we understand the intent behind, the the the contingent sourcing agent and what you're trying to drive with it. Right? So, John, of course, you know, what are the building or what are the building blocks for you to be able to enable, something like this? Right? So at the back end, we understand that there are a lot of foundation elements that need to come together. So how are you thinking and approaching some of this to be able to operationalize, operationalize an agent like this? Yeah. Absolutely. So to deliver on this vision, we're really focused, on all the elements I touched on earlier, but let me let me recap those through here again. So first foundation is your data. So what insights do you need, and how do you organize that data to provide those insights? So, how do we enhance our API suite to ensure that data can flow in a way that it needs to to help power those Agenca capabilities? This is where we're gonna be investing a lot of our time over the next year. Next is the interface. How do users engage with the AI? This is where conversational experiences come into play and really helps paint that picture as users start expressing their intent through natural language versus following, you know, a click by click step through workflow, as they do today. So we plan on introducing a conversational experience into Venly so that users will be able to do more and more through that natural language engagement, which is really becoming the the norm of using applications today. And finally, as we add more more of the specialized agents like CSA that we looked at as a sourcing agent, the orchestration model, the system becomes more and more autonomous, keeping an eye on activities for you, monitoring compliance like we talked about, and really flagging those risks and decision points. So with that investment in AI, our focus remains exactly how it always has always been with that work at Bentley and really solving these core problems that our customers face and helping them manage their extended workforce. So how do we accelerate that time to value? How do we achieve true total workforce management? And how do you optimize a complex global program? So our commitment to solving these problems, it hasn't wavered. What's really changing today is our toolkit. New technologies open the door for us to really think about how we solve these, business challenges, but with a a new lens and a new way. And, John, in that final piece, you mentioned that, you know, we're still focused on solving the challenges that our customers face and achieving that strategic move to total workforce management has been on the agenda for many companies for a long time. So, Krishna, this leads us into our next piece here. If we could focus, on, you know, the that cover, how does agentic AI accelerate the long promised vision of total workforce, total talent management? Yes. I think, Catherine, I'm really cancel whenever I talk about total talent. Right? Because I think for years, total total talent or total workforce acquisition, total workforce management, This has mostly been a buzzword, and, of course, the reason is because systems couldn't really talk to each other. However, with AI, now that is changing. Right? So AI is that missing piece that can essentially enable this interoperability and integration between different systems and enable the transition towards a total talent or total workforce orchestration, mandate for organizations. So of course, there are different levels of maturity through which organizations need to jump through, to to be able to realize this vision of total workforce acquisition or total workforce management. So the way we look at it at Everest Group when we think about the maturity, so there are different stages of total talent maturity. Right? So you start off with, let's say, total talent visibility as a first step. So you get full visibility into your permanent and contingent workforce as a first step. Really understand your workforce mix, the workforce dynamics, and what and how is that shift really changing year over year depending on market situation. So these are all of this becomes really like an important input, and the visibility elements sort of acts as a foundational lever to be able to transition you to more advanced stages of total talent. Right? So next up, we get into things like total talent acquisition. So purely from a recruitment and hiring standpoint, you can with the visibility that you now have, you can make better workforce planning decisions. You can, think about diversifying your talent channels to be able to truly unlock the potential of various channels that they can offer and unlock access and proximity towards those skill sets, across both permanent and contingent workforce channels as well. Right? And then, of course, gradually, then you move towards total workforce management. So that's going to be the ultimate or or or the nirvana that we want to achieve. Right? Where you're looking at all of your HR and talent related operations more holistically, right, through the total talent lens as well. So all in all, so this is how we look at the evolution of the total total talent landscape. However, I think when it comes down to what really enables this, as we discussed earlier, first and foremost, getting visibility into that data or your total workforce is a first step. So that's where you need to have a strong data foundation. So data analytics become really important. So you need to have that integrated workforce analytics across both permanent and contingent operations. And if you look at some of the statistics here, they can slightly be disappointing on the face of it because when we talk to contingent workforce leaders, 48% of them said their programs are still stuck in basic hindsight analytics. Right? Looking at what happened, let's say last quarter or the last year. So only eight percent have really advanced to predictive analytics, which is embedded and only 12% utilize peer benchmarking effectively. So bringing that external market data and intelligence for better workforce planning and then holistically approaching this becomes really important. So most enterprise programs are running on hindsight when businesses, and, of course, you know, they're running on data, and they're not really looking at predictive foresight because they don't still have that underlying data infrastructure in place. So, Catherine, if you move to the next page. Right? So what happens now, right, with AI? Clearly, VMS is now becoming that key cog, right, in the entire wheel for us to be able to operationalize because VMS is acting as that system of record or the system of proof for contingent workforce data. So essentially, everything that's contingent can be tracked and managed using a VMS. So once it can seamlessly integrate with other enterprise applications on the permanent workforce or the full time workforce side, so you are getting that holistic visibility into your overall workforce. So essentially, agent AI now changes the game altogether by shifting the paradigm from worker category view to work diction view. Right? So instead of starting with arbitrary silos, like, for example, should a FTE or a contractor or SOW execute, a particular piece of work? So essentially, we're still looking at workers and the workers that would be required for or or or the workers that you should be hiring for a particular role. From there, you are actually going to shift to, let's say, the work decision view. So I said to think about how the work can be accomplished. And it's not really looking at, hiring a particular type of worker, but it's more about how to get the work done by breaking down works into tasks and then eventually deciding what is the right way to go about hiring or executing that particular task. Right? So a few examples that we have outlined here. So essentially, you start thinking along the lines of what specific skill sets are required, how urgently should something be delivered? And what is the cost or risk or compliance trade off across the various talent channels that you have, that you are operating? So all of these questions become central, and that's the transformation that AI is going to drive you towards. Right? So you're not really thinking about only a permanent worker or a contingent worker. You are clearly asking the right questions around how to get the right the access to the right skills, how to get the work delivered on time at the highest quality, and at at what trade offs. So AI clearly bridges these silos by evaluating all worker pools across internal employees, external contractors, your gig talent, and AI agents, of course. And I think the AI agents part of it is also clear, clearly important because workforce now is not just about permanent or contingent workforce alone. So your AI agents are also an important component of the entire workforce mix. Right? So we should be thinking about what should be the right mix of full time employees, your contingent workers, and AI agents as well. So I know I think there is a lot to unpack here and lot digest here. So of course, Total Talent, like I said, it's not straightforward topic, but clearly we have the right ingredients in place, especially because of AI that, that we are getting closer to be, to being able to operationalize it and of course, operationalize it at scale as well. So John, again, coming back to you. So what do you think about thinking about this? Right? So does this really align with how you are operationalizing or how you are thinking about total talent in general? Yeah. Definitely. That one of the terms you mentioned that work decision view, is something we hear a lot, and it's it's often talked about as kind of this this theoretical dream. Right? But when you when you have your HCM system and your VMS natively connected, it doesn't have to be a dream. It can actually be an operational reality for you. So on the screen, we have the blueprint, for what we consider the next generation of total workforce management. So at the center, you have Workday HCM for your employees, and you have Workday Venly for your extended workforce. And then Workday has recently announced, our new Workday agent system of record, which really governed that digital labor. So orchestrating all this for us, Asana, which is Workday's next generation AI core, and it's all gonna be surrounded by an open API ecosystem. So this is how, enterprises are gonna be able to achieve that full visibility, unified governance, and the seamless orchestration across, their entire global workforce. Perfect. Thanks, John. So I think, you know, for organizations really looking to make that move to total workforce management, Workday has all the the aspects needed to connect those different worker populations, and there is no doubt that AI is core to our company strategy as we do move move forward. So this has been a fantastic discussion. It's flown by. So a quick reminder that we do have the q and a the q and a box there. So you can submit your questions via that, and we'll get them to Krishna and John. So while people are getting in their questions, we'll just quickly recap on the session today. So first, if you want to leverage AI, you have to get your data house in order. You simply can't build predictive programs. You know, if your full time employee and external worker data is stuck in disconnected silos. Second, the role of the VMS is fundamentally changing. We are moving beyond the system of record of today, and the future is a proactive system of action that works right alongside your teams to anticipate bottlenecks and orchestrate the right next steps. And lastly, achieving, a true total workforce management strategy means breaking down the historical walls between your permanent employees, extended workforce, and now, of course, your your digital workers. And connecting your entire talent pool, you know, on a unified platform is the only way to treat skills as your primary currency and drive that real business agility. So we'll take a look at what's coming in via the q and a. So do do, Krishna, if we if we start with you, you know, based on your conversations with buyers for the peak matrix, do you think enterprises are actually ready to let AI agents make decisions, or is there still a lot of hesitation around compliance and risk in that area? Yes. I think, Catherine, the short answer, of course, is there are still, concerns and there's still a lot of hesitation. But I think I would still refrain from answering with a simple yes or no. Right? Or or rather, it's it's not as direct as we can as one would imagine. Right? Because ultimately, it depends on a lot of factors. There are organizations with a certain level of maturity that are able to actually, rely on AI and essentially leverage agentic AI for more autonomous decision making. However, there are others which are still some distance away from being able to fully use autonomous agents. So having said that, I think we did discuss about the human in the loop component earlier. So I think the human in loop is going to be critical. It's going to be important. So it becomes important for organizations to really determine what should be the steps or actions that they can rely on AI for, and what should still be those actions or steps that should be, the responsibility of humans. Right? So if I think about a few examples, I should do things like checking compliance, identifying exceptions, or recommending actions. So essentially looking at things like things which are reversible in some ways, which are with some sort of predefined which are working against some sort of predefined policies. So these are the steps where an organization is still open to leveraging agentic AI and rely on agentic AI to autonomously execute some of these actions. However, when it comes to things like, let's say, compensation, right, or even worker classification, supplier selection. So all of this, I think the human related component, the human involvement is still on the higher side. So, again, in a nutshell, no simple answer here. So it ultimately depends on how ready you are and what your maturity is. So you need to get an understanding of your current maturity and practices before you determine what is the optimal level of agent KI enablement within your talent operations. Yeah. Perfect. And, John, there's some questions here. They look specific to the contingent sourcing agents, so maybe we'll take some of those. Does the AI sit on top of the VMS, or is it integrated into the VMS? Yeah. A little bit of both. I think our general model is obviously, you know, what can we bring that's embedded directly into the tool, but we also understand that the AI ecosystem is quickly evolving. Right? So you may be using other tools kind of over the top, and, really, that consolidated experience is what you're after for your users. And so that's why when I was talking through a little bit of a lot of the near term, a lot of it's gonna be API based so that we can expose that information into other services if needed, so over the top. And then, obviously, the information directly in Bentley, like the kind of sourcing agent that we looked at earlier. Perfect. And then there's another one here, which maybe both of you can chime in on potentially. What data or operating conditions do companies need in place, I suppose before AI can effectively work within their contingent workforce program. I think I'd I'll take first swing, Krishna, and you can hop in there. So, I think. it's it the thing is we talk to customers that, it it can be this this big mountain, you know, this big hill to climb. And so typically advised to just get some of the foundational data out the door. Right? Focus in that area. At least then if you're pulling off on reports, you're doing other things maybe outside the system, you at least have that grain where you can start to compare apples to apples. And I really think that's the first step. There's a lot of things you can incrementally build into over time, but none of those are going to work. And all this agentic orchestration and these other things we've talked about that sound really great are only gonna be as good as having that foundational data aligned. If you don't have that foundational data aligned, you know, how can you match your employees, your nonemployee, and your agents and really have that full conversation? Yeah. Absolutely. I think, John, the only point I'll add here is, of course, you you did talk about some of the foundational data. You're absolutely right. So, essentially, all the data that sits sits across your systems of record across both full time permanent workforce and continued workforce, that would definitely become the first step or the first thing that you need to sort out. And then I think the other key thing to call out here is the access to external market data and intelligence as well. Right? Because we are not simply looking at AI here because we're not looking at predictive and prescriptive analytics based on just historical data alone. So when you think about execution or executing actions at scale, you also need that external market intelligence because the external context becomes important along with your organizational context. So you need to have both of these elements sorted out and together to be able to really, achieve agent decoquestration. Perfect. And there's another one here on the contingent sourcing agent. So what specifically is AI referencing to determine the the required versus preferred skills for a position? Sure. So we're pulling a lot of that directly off of the job description as it's created out. And then I've seen a couple other here on bias testing and things along those lines. So we'll try to pull some of those in as well. So I talked a little early with customers now, and and we'll be generally available, later in the year. And a lot of this is going through. We obviously wanna make sure that, we're not introducing bias into the systems we're going through this. So, luckily, we we have a whole lot of recruiting data at Workday we can take a look at, and use that as a lot of the, as we're going through this evaluation. So as you're as you're analyzing the case of sourcing agent, please reach out. We can give you a lot more information about how, I want to adjust it on here on how we went about that testing and and kind of the standards we're going against, in order to, as best we can, help ensure that, you know, there's no bias introduced into, into the system. I saw another question about overriding the candidate grading. I'll say most things that Vendly does is there's gonna be a permission tied to it. And so you can kinda decide maybe you keep these grades only on you know, only the back offices, and you continue to shortlist today. There's a permission around being able to see and utilize a lot of these items so you can kinda determine what role, in your in your program is gonna actually take those actions. Perfect. And there's a maybe a broader one here. Given that many candidates oh, is it gone? Oh, sorry. It just moved in. They are now using AI to align their resumes to job descriptions. What are your thoughts on how managers slash MSPs can better screen? That's. I think I think, real go for it. That's that's that's that's an interesting one. yeah, I I think the interesting thing here is if you look at, how candidates are being assessed or evaluated. So I think clearly organizations are understanding some of the challenges that come with AI and how, candidates are trying to bypass some of this, right, and use it to their advantage. So I think that's a place where the assessment approach of organizations is also changing. Right? So we're seeing a lot more focus on, not just resume based screening or not just face to face interviews even before you get into the highly managed interview stage. So within the screening process itself, candidates are expected to take tests. Right? And these tests, need not necessarily be, let's say, a simple static website based test. Right? So these can be more interactive tests. And, of course, there's a great deal of proctoring in place as well where candidates are really evaluated whether they are addressing or answering this test, in the way they are intended to without really leveraging any of the external tools to assist them with their responses and so on and so forth. So I think the, AI readiness and AI related assessments are also becoming an important part of the entire evaluation process, which is in some ways also helping organizations to minimize this, but, of course, no. I think clearly that's, that's a challenge that we are still tackling, and I think we'll not have any quick fixes or answers anytime soon. Yeah. Exactly. And and and there's another one here then that's I suppose could be broad, but also maybe in the context of the technology itself. And to clarify, what are what are we including when we talk about compliance? What are we thinking about there? In compliance, in in I guess, I can't really ask a clarifying questions, can I? So when looking at compliance, I'd say there's a lot of different angles, to look at. Even, you know, Krishna was talking a little bit about, you know, during that interview and screening, really, like, are you doing the things that you're you're putting on your resume as one area of compliance? Obviously, there's compliance when it comes to tenure policies that can vary by country. So it really it really just kinda depends which lens and which angle. But even inside those items, the more the more data you have, the more you can you can put together, which is why it's a lot of the initial focus is on making sure we're getting that data aligned. Yeah. And there's another one here around what are the most common data integrity issues you see that make it challenging for an organization to implement AI in contingent workforce management? I'd probably answer it in two different ways. Here, we obviously talked a lot about total workforce, and I know I already answered the one question on kind of your your your data, foundational data. And I think that's really critical when you start looking through, any type of predictive analytics or other agents' experiences that are really looking at the whole the whole workforce. I would say a a planning function is not just contingent. It's not just employee. You really have to look at that whole picture together, and the only way you're gonna do that is is having your foundational data aligned in. Some other things that I think that are gonna be interesting here, is better better data validation. We hear a lot from from managers, from others, and they're using a lot of systems and a lot of tools. And I think that's where some of that conversational experience and natural language, work is going to help. If you got somebody who opens a job three times a year, I think getting it right out of the gate is gonna help a lot of things along the way, and I think that's where natural language, conversational experience experience is really gonna help shine for us. Anything you'd add to that, Krishna? Yeah. I think I think in terms of data integrity, perhaps, I think the the thing that comes to mind is still, a lot of organizations tend to deal with lot of inconsistencies when it comes to the job role or skills taxonomies. And especially in this current day and age where the skill sets that you require within your organization, how quickly they are changing. I mean, think thinking about future proofing your workforce, the skill sets that you would require for tomorrow, not having that skills documented, not having a skills taxonomy in place, which can which is being also updated real time. Right? That tends to become become a challenge. And, of course, the reason for this is, generally, if you look at workforce management programs or talent programs, there are plethora of tools that all organizations are operating. Right? So some of it still continues to be manual. Some of this data sits across multiple systems, and we did talk about that fragmented experience with technology and, you know, data not seamlessly being consolidated in one place, which becomes a challenge. So when we talk about data integrity and lack of, proper visibility, so we're looking at some of these elements as well. Perfect. And I'm just looking at the time here. I know we're kinda coming up to the the two minute mark, so I'm gonna move on to some of the the final wrap up piece pieces here. So, we strongly encourage everyone attending today to scan the QR code on screen and download Everest Group's 2026 VMS peak matrix report. So that will help you learn more about, you know, what's going on in the market and across VMS providers overall. We also invite you to connect with a Workday Vanleigh expert to to assess your program's AI readiness. And then we want to highlight a couple of upcoming webinars here as well, so please sign up for those if if any of those are of interest. And if you're in the Chicago area, you can join us for an in person event there with Workday Elevate on August 6. And and last but not least, registration is also open for Workday Rising, so that's taking place in Las Vegas on October twelfth to fifteenth this year. So that's the perfect opportunity to connect with peers and, of course, the the Venley team on-site there. And then finally, Krishna, John, thank you so much for sharing your expertise and vision today, and a huge thank you to everybody online who's who's spent the time with us as well. As a reminder, if you do have a few moments, please do fill out the feedback survey, and, you know, that can help us improve these sessions going forward. And I think with that, we'll wrap it up. And everyone, you know, have a have a wonderful rest of day. Thank you. Thank you. Thanks, Roman.