Dreamforce 2025 | Greg Shewmaker, r.Potential & Denis Machuel, The Adecco Group

Greg Shewmaker, r.Potential & Denis Machuel, The Adecco Group

Exploring the Role of AI and Digital Workers in Modern Enterprises Denis Machuel of The Adecco Group and Greg Shewmaker of r.Potential join theCUBE at Dreamforce 2025 to discuss the evolving landscape of artificial intelligence and digital workers. The conversation, hosted by Dave Vellante and George Gilbert of theCUBE Research, examines how r.Potential, emerging from Adecco in collaboration with Salesforce Ventures, leads in digital workforce transformation. Shewmaker shares insights into the creation and vision of r.Potential, aiming to harness AI's capabilities to enhance human potential. They emphasize how r.Potential fosters a symbiotic relationship between digital and human workers, ensuring AI supports rather than replaces human effort. The discussion also underscores how theCUBE Research offers critical narrative insights. Machuel highlights that the initiative ultimately seeks to integrate human and digital workforce dynamics seamlessly, providing organizations with enhanced versatility and productivity through AI. Key takeaways include the strategic partnership with Salesforce to scale AI-driven solutions, building digital workers that earn trust and can autonomously manage business functions over time. Machuel notes that, based on gathered insights, the focus should be on creating a skill-based operating environment rather than a role-based one, aligning AI capabilities with organizations' strategic goals. The video offers an in-depth look at aligning AI within corporate ethos, ensuring supportive infrastructure for digital and human collaborations.

Share this session

Clips
More from Dreamforce 2025

Greg Shewmaker

CEO

r.Potential

Denis Machuel

CEO

The Adecco Group

play_circle_outline Empowering Human Potential: r.Potential's Journey as a Salesforce Portfolio Company Leveraging AI for Transformative Workforce Solutions

play_circle_outline Rationale behind the spin-out from Adecco for innovation in workforce changes.

play_circle_outline Empowering Collaboration: The Role of Digital Workers in Enhancing Human Capabilities and Fostering Symbiotic Relationships

play_circle_outline Chief potential officer role designed to assist CEOs and accelerate decision-making.

Info
Transcript

Greg Shewmaker, r.Potential & Denis Machuel, The Adecco Group

Greg Shewmaker

CEO r.Potential

Denis Machuel

CEO The Adecco Group

In this Dreamforce segment, theCUBE’s Dave Vellante and George Gilbert sit down with Greg Shewmaker, chief executive officer of r.Potential, and Denis Machuel, chief executive officer of The Adecco Group, to explore what it really takes to bring digital workers into the enterprise. The discussion traces r.Potential’s origin as a spin-out (backed by Salesforce Ventures and Adecco) and unpacks the company’s thesis: start with humans, not the tech. The guests detail why “digital workers” must earn trust over time with persistent identity, memory and auditability... Read more

explore Keep Exploring

What is the reason behind the founding of the company and the origin of its name? add

What strategy did the company implement to adapt to the changes brought about by AI in the workforce? add

How does managing both human and digital workers change management practices? add

What attributes are being developed for the chief potential officer as it evolves in automation and human collaboration? add

bolt Powered by CUBE AI

Greg Shewmaker, r.Potential & Denis Machuel, The Adecco Group

search

Dave Vellante

>> Hi everybody. Welcome back to San Francisco. We're here in Moscone South at Dreamforce 2025. Winding down day three. I'm Dave Vellante with my co-host George Gilbert. We're taking over John Furrier and Gemma Allen who had to go back to New York City at an NYSE studio. We're really excited to have Greg Shoemaker here. He's the CEO of r.Potential, pretty interesting name. We're going to talk about that. Denis Machuel, who's the CEO of The Adecco Group. Good to see you, gentlemen. Thanks for coming in.

Denis Machuel

>> Hi Dave. Hi, George.

Dave Vellante

>> Hi guys. Thanks for having us.

Denis Machuel

>> Thanks for having us.

Dave Vellante

>> So r.Potential just launched this year, I believe, right?

>> Yeah. March.

Dave Vellante

>> You're a Salesforce portfolio company, Salesforce Ventures, and a spin-out of Adecco, correct?

>> That's correct.

Dave Vellante

>> So tell us more. Why did you guys start the company and where'd you get that cool name?

>> Well, the name comes from, the whole idea is to strengthen and secure the full potential of humans with AI. Humans as the sort of main actor, AI as a supporting cast. So that was the idea, and we had thought about that two years ago before anyone was even talking about agents. We're like, we see this coming. We know it's going to change the workforce. We don't know how yet. But what we do know is that we need to start sort of deliberately creating new ways to protect and ultimately elevate humans in this whole thing. AI was going to have a profound effect on the workforce, there's no doubt about that. But we needed to start with the human, not start with the technology.

Dave Vellante

>> And what was the rationale of the spin-out from Adecco? You guys, obviously well-known, I think Zurich-based right?

Denis Machuel

>> Yeah. .

Dave Vellante

>> And public company and very successful. What was the thinking behind the spin-out?

Denis Machuel

>> I think there were two things. The first is since inception, our role, our business has been to match the world of resources with the world of work, with a business, and to ensure that our clients would have the right resources to deliver their business goal. And when you talk about right resources in the times of AI, you understand that the magnitude of change that is necessary to have agents and digital workers come into the workforce is so big that you have to do something differently. And actually, it all started by what we did with Salesforce. We did a killer company workshop to reinvent how we could disrupt ourselves. And we came up with this concept of r.Potential. And the best way to disrupt ourselves was to launch that company outside, even though we are the shareholder together with Salesforce, to relaunch that rocket to go as fast as it could to help us create that new horizon that is going to be needed for our clients as they reinvent the organization of work.

Dave Vellante

>> Well, George, when we had Mark on our program, it was like the week that he wrote that Wall Street Journal article saying, we'll be the last generation of managers to manage only humans. So I can see why he would, I presume he was involved in the-

George Gilbert

>> 100%. That's right.

Dave Vellante

>> In the investment. I can see why he would be attracted to that.

George Gilbert

>> yeah.

Dave Vellante

>> But so the question at the keynote was always, well, what's the vision? What's your vision for the future? So what is the vision that attracted?

Denis Machuel

>> I think there's kind of two dimensions to the vision. The first one is how do you build relationships between human machines that don't exist today? And so that was what Denis was mentioning earlier. We were thinking about it as true digital workers, right? Because if you think about AI in the terms of software, well, software is a product. So software, you can return a product if it doesn't work, you can sunset a system, you can fix the system. But AI introduces a whole new category of risk and responsibility. We know it hallucinates, we know it has biases, it gives unpredictable answers. Well, guess what? So do humans. And so if you think about AI much more as a worker versus software, I think that's starting off in the right path. So if you can build these human machine relationships, and then if you can scale that through coordination across the enterprise, that's where we talk about organizational potential. But if you can do those things, they're always going to be in conflict with one another. And I think that's where the problems start.

George Gilbert

>> What are the ingredients to make that digital worker and then to scale that across the enterprise? We were talking earlier about like you mentioned, memory I think was one of the things. An agent inside its little non-beating heart, there's an LLM, but there's a whole lot of other stuff around it. And then if you want a team or even an army of agents, you need, it's not just scaffolding. You need a platform.

>> Yeah, that's right.

George Gilbert

>> Tell us more what the pieces you need to fill in.

Denis Machuel

>> So the way we define a digital worker is that it's AI that earns trust over time. And as it earns trust over time, it becomes sort of capable of having more autonomy over business functions. Again, just like a human being does. You don't hire an intern and put them in charge of finance on day one, right? They have to earn their way towards that. And we believe digital workers have to be the same. So an agent is a smart tool. It can do three or four or five tasks, and ultimately it'll do more and more. But it's sort of ephemeral. It has short-term memory. So once you're done with it's kind of done. Whereas a digital worker, it has persistent memory, it has an identity that it maintains. So a digital worker doesn't try to do everything like a general LLM says, this is what I am tasked to do, this is what I'm accountable to do. And it could be held accountable to do that. It could be audited, it could be accredited, and it has a role in the organization. And so this is, we're building towards this. It doesn't exist today, but we think that someone has to set those standards or otherwise there will be no standards.

George Gilbert

>> So on that topic though.

>> yeah.

George Gilbert

>> We saw the, I think they call it observability that captures all the reasoning traces and the data that's fed in context-

>> Yeah, that's right.

George Gilbert

>> And the tool calls. It's almost like the new big data. We used to have click streams. Now we have the agent exhaust.

George Gilbert

>> That's right.

George Gilbert

>> You could build memory on that or do you abstract and synthesize that so that you're not essentially stuffing a whole amount of raw exhaust. How do you build memory, and then how do you build provenance for the agent so that it can be audited, things like that?

George Gilbert

>> Yeah, so that's a good question. So we're thinking about memory sort of on two dimensions. Again, it's this persistent identity-bound memory. So just like humans, we see and hear everything, but we sort of put into our memory the things that are most important. So how do you train a digital worker to do that? What's relevant? So not just what happened in this last session, but what happened a week ago, two weeks ago, three months ago? How do you start to maintain that context over time? Right? That's super important. But when you talk about observability, that's sort of the technical element of a digital worker or an agent. But what about the business context too? You still have to have that as well. So you have to look at these digital workers on both the technical sort of performance, but also the business performance too. So that's the part we're thinking about is great, MuleSoft and other technologies that we see here have that technical observability, but we have to bring to it the business part of it.

George Gilbert

>> Does that mean you're distilling what's relevant that was captured raw? Or do you pull in additional context so that you're building a different type of coherence?

>> Yeah, so one of the things that is what we built above the LLM is sort of this four-component sort of orchestration layer or this intelligence layer. And one of the components of that is the context layer. And so we bring context in from within the organization, but also from the external world. And so this is digital workers constantly learning from different sources, both the interaction it's having with its human counterpart, but also a broader world.

George Gilbert

>> So this is... Just one last question. This is really interesting, because what used to be Data Cloud, Data 360, I was about to say 365, but that's the other that's render.

>> That's right. The other one.

Dave Vellante

>> They're connecting to that though.

George Gilbert

>> It is supposed to be the context for agents, but you're talking about a different type of context. And you're also talking about something that I don't think is in Data 360 yet, which is persistent to an individual agent.

>> That's right. That's exactly right. But we are working with the teams, the Data Cloud team.

George Gilbert

>> So it might expand?

George Gilbert

>> It could expand, that's absolutely.

George Gilbert

>> Okay. Okay.

>> So we feel like, and I think Salesforce would agree with this, we're sort of on the cutting edge beyond even where their particle maps are today. And we're feeding back to them some things that could be interesting for them to build at scale. And that's beautiful about the partnership is that we're sort of pushing the boundaries and then they have the enterprise scale, because that's right on.

George Gilbert

>> The identity model, I've heard a little bit like the Sequoia guys talking about this.

>> yeah.

George Gilbert

>> A very different identity model when it's one thing for the agent to essentially borrow your credentials. And it's another thing for the agent to have its own credentials.

>> That's right. So the agent has to have its own.

Dave Vellante

>> I think they're about over 30,000 employees at Adecco. It's a-

Denis Machuel

>> . We have more than 100,000, yes.

Dave Vellante

>> Okay.

Denis Machuel

>> And a million people on our payroll through all our temporary workers that we put to work every

Dave Vellante

>> Okay. So you got 100,000 people. So ever since Mark came on, he wrote that article, I asked him, I've asked all the CEOs, what does that mean? If you're going to be managing not just humans, but digital workers and humans, how does that change the way in which you manage?

Denis Machuel

>> Well, let me get back to our purpose. Our purpose is to make the future work for everyone. And everything that we're building here is about making sure that humans are not being replaced, but augmented. That we create the future that we want to see happening. That means a future that is bound to create a human-centric AI workplace. And for that, we believe that what we have to achieve is to create these digital workers that work in symbiosis with the human workers. And as I said earlier, as we provide resources to our clients, the proper resources, I think eventually we're going to be able to provide the combination of the human workers and the digital workers because we are creating, as Greg was saying, the operating system for them to work together. And that's the power of what we're creating with r.Potential is the standard of digital workers, but also the way they're going to interact with the human world.

Dave Vellante

>> So I'll tell you that answer is a combination of what Mark laid out in a vision and what Arvind Krishna told me when I asked him that question was essentially our job is to give the tools to the humans so that they can be more productive. But something that Greg said sparked, because I love this question. What does that mean? How is a pretty profound change? You talked about, Greg, earning trust.

>> Yeah.

Dave Vellante

>> And immediately I went to the Amazon leadership principles. And so I was thinking about Erik Brynjolfsson, the economist who was at the time at MIT, now he's at Stanford. Is that a step-up? And so-

George Gilbert

>> That's a debate you.

Dave Vellante

>> So he wrote a book with Andy McAfee called The Second Machine Age, , and we did a CUBE thing with them. And we went through what are humans able to do that machines can't? I mean, at the time they couldn't even climb stairs and robots can do that now. And then we just recently saw him again and so much has changed. But I'm looking at the Amazon leadership principles. One of them is earned trust.

>> Yeah.

Dave Vellante

>> The reason I bring this up, kind of a long-winded circuitous question, but there are certain things that you want agents to do, earn trust, but you don't necessarily want them to be curious, which is another Amazon principle. That's something that a human could do. Maybe customer obsession is something that they could do within their little zone. So my question to you is, how do you see the evolution of the training of those agents? What are those sort of attributes that you want them to build upon as they progress?

George Gilbert

>> So just to give a real concrete example, so the very first digital worker that we're building is called the chief potential officer, and it's meant to be a digital colleague of CEOs first, and this is Customer zero right here, Denis, the chief potential officer. I can tell you, because we're only six months old as a company, when we first started producing some of the outcomes from the chief potential officer, it was maximum 10% automated and 90% human. Because of your thing. We didn't trust it. We had to say, hey, here's some answers. And some of them are good and some of them are, et cetera. The human had to get involved and say, "Okay, wait, this is completely screwed up and wrong." And then we sort of evolved to 30% and 50%, et cetera. Today we think it's probably about 75% automated in terms of what this chief potential officer or CPO sort of produces, but they're still human there to sort of make sure and maintain that trust. By January, we're hope we're above 90%, but I think that 90 to 100% is going to be a much longer period before we fully trust these digital workers. And so that's what I mean by earning trust is the human has to be in the loop. It has to be in control. It still makes the decisions. These things can process a tremendous amount of context and information that we could never do, but at the end of the day, it can't make the decision for us.

Dave Vellante

>> And what I'm inferring is you're going to be supercharging these cultural edicts or first principles and bringing in an assistant to the human so that they can be think bigger or be right more or be more curious or be more customer obsessed.

Dave Vellante

>> That's exactly right.

Denis Machuel

>> Yeah. Yeah. And if you think at the chief potential officer-

Dave Vellante

>> I love it.

Denis Machuel

>> Which is the first product of r.Potential as the next member of the executive committee, as the buddy to the CEO, it's a safe buddy. Because you can interact in many ways without feeling that your judge or whatever you say can be interpreted. That's the first thing. Second thing, it bring different angles, and that's what you want from your exec team. Bring the new ideas, bring different angles. And because the chief potential officer is plugged into all the external data that we've loaded on the platform, but also the internal data of the company, it has a unique perspective on what's happening and brings you good angles. And then, again, as Greg said, I'm going to be the one making the decision, but the capacity to accelerate execution going from insight into execution because then the chief potential officer interacts with the broader ecosystem, interacts with colleagues, brings people together, orchestrate conversations, et cetera. That's a way to accelerate execution.

Dave Vellante

>> I love that. And I want to follow up, and I think it's an obvious answer, but I want to hear it from you. Your buddy, I talk to my LLMs all the time, they're actually quite useful. We have wonderful conversations in the car, but we're talking enterprise here. This is different. So presumably there's a proprietary data set that is informing that I'm talking to. I wonder if you could elaborate on that.

Denis Machuel

>> Sure. Maybe it's... I mean, of course. I mean, first thing, it's r.Potential is enterprise grade. Because yeah, it's true. There's a moment when you plug the platform into your HIS, HR system into your financial system, into your CRM, et cetera. There's a little bit of an open kimono moment where you believe that this has to be trusted. And also thanks to the partnership with Salesforce and MuleSoft, we have definitely an enterprise grade architecture. And you need that because you need that granularity of insights that helps the digital worker give you the right angle, give you the right perspective, and of course can dig you, you can challenge, but that's necessary, right?

>> Yeah.

Denis Machuel

>> And so we build an architecture that is rock-shotted.

>> Yeah. And that memory and those conversations that the chief potential officer's having with the human executive is maintained in sort of our architecture in our layer. And so it's not going back to the LLMs right now. So this is what we talk about with memory. That memory exists in our platform to maintain that security and that confidentiality so the CEOs can have the confidence to start to share things in an intimate way with-

Dave Vellante

>> Which their proprietary advantage.

>> That's right. Exactly.

Denis Machuel

>> Your strategy is not going to go on the bigger .

Dave Vellante

>> Yeah. Be careful what you tell ChatGPT.

George Gilbert

>> That's right. Exactly.

Denis Machuel

>> That's absolutely true. Absolutely true.

George Gilbert

>> They train on that.

Denis Machuel

>> Yeah, yeah.

George Gilbert

>> I wanted, I usually ask the product questions, but I wanted to ask a big picture question. Modern large-scale organization first started with the Prussian military, the hierarchical organization, and then the org chart really came out of railroads where you had this geographically distributed, but you had to coordinate the work. Now that we are going to have digital labor, what do the org structures look like?

Denis Machuel

>> That's super interesting. It's part of the reinvention that's going to happen in the world of work, in the way work is organized. First of all, we've got to think about the work being now no longer job-based, but skills-based. The is going to be skills-based and not job-based and role-based. It changes. And so that's the first thing. And second thing is things are moving so fast that you have to articulate the agility of the company through, as I said, this symbiosis between the agents or the digital workers and the human workers.

George Gilbert

>> To align them.

Denis Machuel

>> You have to align them. But the org chart is... And first of all, let's be clear, the org chart only tells you a little bit of what's really happening in the company. Because org chart, job descriptions, this is not how work gets done. And the gap between your role and actually you do is massive. And that's the power of that platform being plugged into how the work gets done. That gives you the best possible insights.

George Gilbert

>> Just a quick follow up on this. When you go from function or role to skill, who's orchestrating all the skills in alignment for an end outcome for a business outcome?

Denis Machuel

>> Well, that's, for me, that's the fundamental responsibility of management. And again-

George Gilbert

>> But is management now augmented by digital technology that couldn't really do that sort of alignment work before.

Denis Machuel

>> Because it'll have the proper insights in understanding exactly what's happening. And that's what we've built.

Dave Vellante

>> That's right.

Denis Machuel

>> Do want to?

>> I was going to say, so one of the things that the chief potential officer produces is what we call units of potential. And it's these dynamic configurations of work that don't exist today. So it's what is my desired outcome? What am I actually trying to accomplish? It has to start with that first. Today everything about AI is starting with a tool. It's starting with a hype, Hey, I got a license. I bought something. Now how the hell do I use it? We're saying, no, it's backwards. Let's start with what am I trying to accomplish? What are my constraints that I actually have real world constraints? What should humans be doing in this particular case? And then what is AI actually capable of doing? If you can put those four things together, that becomes sort of the new unit of work. Now it's still experimental. We still have to prove that up, but we think that's how work is going to be done in the future in much more dynamic ways.

Dave Vellante

>> So what are you proving? You raised I think 16 and a half million. How are you deploying that capital? Is primarily engineering? Are you doing any go-to-market at this point other than belly-to-belly or?

Dave Vellante

>> Other than being a Dreamforce? No. It's all engineering. We're building the platform.

Dave Vellante

>> Right. I mean, you got to determine product-market fit and then scale your go-to-market.

Dave Vellante

>> That's exactly right.

Dave Vellante

>> Don't rush to go to market before you have a product that people will buy.

Denis Machuel

>> But the sort of go-to-market we do is to talk to the C-suites about how they think about this reinvention. And that's how we feed into the way we designed the platform. Because as soon as you talk to A CEO or CDIO or CHRO, they're all like... You heard this study from McKinsey, 95% of the AI pilots fail.

Dave Vellante

>> The MIT. Yeah,

Denis Machuel

>> MIT.

>> Yeah.

Denis Machuel

>> And all the reflection is about where do I start? How do I get things done? How do I bring people along? And so all these reflections that we have and the discussions about the concept that we're building feed into the way we create this architecture. So there is no clear go-to-market, but there's a lot of, and there's a lot of interest from our customers-

Dave Vellante

>> I'll bet.

Denis Machuel

>> In terms of how we think about the way we help them organize the future of their organization.

Dave Vellante

>> And Adecco as customer Zero increases the probability that you'll succeed.

Denis Machuel

>> That's right. That's right.

Dave Vellante

>> As you're real-time feedback from a trusted partner.

>> That's exactly right.

Dave Vellante

>> It actually has an observation space that is pretty deep and wide.

Denis Machuel

>> Well, that's the point. I think not only is it a massive global organization, but it does workforce management for every major company in the world. So it understands the world of work better than probably any company out there. And so if we can have that perspective to start with, that's going to help us build a standard. And we really firmly believe that someone has to build a standard for this, otherwise it's going to be left to the open market and everyone's going to define it themselves. And who's going to pay the price? Is going to be the worker.

Dave Vellante

>> So next year, Dreamforce 2026, what do you want to be able to say that you can't say now?

Denis Machuel

>> A high percentage of the Fortune 1000 has a chief potential officer in the C-suite helping the CEOs navigate the hardest workforce decisions of their careers.

Dave Vellante

>> Wow. A year from now? That's ambitious.

>> I think so.

Dave Vellante

>> All right. Well, we're moving fast in this AI age.

Dave Vellante

>> That's right.

Denis Machuel

>> We'll be there. We'll be there.

Dave Vellante

>> Guys, congratulations. And thank you for sharing your story here.

Denis Machuel

>> Thank you. Thank you, Dave. Thank you, George.

Dave Vellante

>> Really excited. Love to have you back next year.

Denis Machuel

>> Love to be back. Thanks guys.

Denis Machuel

>> Yeah. Thank you.

Dave Vellante

>> Thanks. Thank you.

Dave Vellante

>> Okay. And thank you for watching this to Dave Vellante for George Gilbert. We're here winding down Dreamforce 2025. People are sort of exiting the building, but we're here continuing. We'll be right back on theCUBE right after this short break.