Joe Fuller: It almost goes without saying that the pace of change has made the traditional siloed organization uncompetitive. When your workforce strategy hinges on your technology roadmap and your product plans depend on your skills inventory, coordination is essential. Yet few companies align their people management and culture with product development and operations. That disconnect limits agility and the capacity to innovate.

Welcome to the Managing the Future of Work podcast from Harvard Business School. I’m your host, Harvard Business School professor and nonresident senior fellow at the American Enterprise Institute, Joe Fuller. I’m excited to welcome Cisco Systems Francine Katsoudas, executive vice president and chief people, policy, and purpose officer, with her colleague Jeetu Patel, president and chief product officer of Cisco. The pair have forged a productive partnership guiding the firm’s workforce and product strategies through the Covid crisis and the ongoing generative AI boom. We’ll talk about how their groups collaborate and cross-pollinate. We’ll consider how Cisco equips employees to experiment with emerging technologies and develop the skills they’ll need to remain competitive. And we’ll talk about the increasing importance of social skills—in communications, collaborations, and in stimulating constructive debate. We’ll also look at how AI is redefining entry-level work and early career opportunities, as well as reinforcing the importance of onboarding. Finally, we’ll examine the role of leadership in managing sweeping change. Welcome to the Managing the Future of Work podcast, Fran and Jeetu.

Jeetu Patel: Thank you for having us.

Fuller: Perhaps you could just briefly describe each of your respective roles at Cisco to set the stage for our listeners.

Francine Katsoudas: Go ahead, Jeetu.

Patel: I’m Jeetu Patel, I’m president and chief product officer. I run products within Cisco. I’ve been here for about five years. In fact, one of my first conversations I had was with Fran, and it’s one of the most memorable interviews I’ve had. And so it’s a pleasure to be working with her for the past five years.

Katsoudas: Thanks, Jeetu. So I’m to blame for Jeetu being here. I’m Fran Katsoudas, and I lead people, policy, and purpose at Cisco. And what that means is that there’s organizations like government affairs, real estate, sustainability, our people organization. And we believe that when we put our purpose into practice, we drive growth for the business, for our people and for communities.

Fuller: So it’s unusual for us to have two executives from the same company on a podcast. But we’re keen to do that given the way the Venn diagram seems to work at Cisco on the intersections and overlaps between your responsibilities. Could you tell us a little bit more about that?

Katsoudas: So it’s really interesting. I think that it’s probably the way that it should have been all along, but I think at this moment there’s this recognition that people and technology and innovation have to be one and the same and that you can’t look at these things in separate lanes. And so what that means is that Jeetu and I are coming together on a regular basis to really talk about the intersection of innovation and what that means, then, from a leadership perspective, what that means from a broader people perspective, how we need to change in some cases the culture of the company to reach the goals that we have. And it means that, in some cases, we blur lines between our roles in service of what’s best for our people.

Patel: If you think about the fact that Fran owns not just the people side of the house, but the policy and the purpose side of the house, there’s so many overlaps of that, then, with the product side that it would almost be impossible not to work together hand in hand. One of the things that really also was instrumental in making this happen was during Covid when I joined. We hadn’t met for the first nine months, and we had to actually establish a culture in the organization where innovation couldn’t have slowed down while we were still working remotely with each other. And that really helped us make sure that the policy and the culture and the people side was tied to the way that we were going to go run a product organization to make sure that you build the best products in the market.

Fuller: It’s interesting because, despite what I think is a pretty compelling argument for a very close alignment of the human assets function—and I’ll just call broadly “operations”—you actually don’t see it in many companies. And, in fact, a lot of the problems we’ve unearthed in our research have to do with the lack of alignment between those two functions—that HR develops policies, runs recruiting, doesn’t talk to the operators very often about how it’s all working out, and then just gets complaints about policies or demand for replacement employees, not even knowing why the previous ones left or why they were terminated. So I think it may be a coming trend.

Patel: And it is one of those things where you don’t really see the huge compounding benefit of it until you experience it once. What it does is, it allows you to create a strategy around talent technology. And if you look at where we are right now, the policy side of the house—with the public sector and the private sector is so important to tie together—that that can’t be something that’s standing on its own on the side that has to be fully integrated with the way in which you think about doing the innovation in the core fabric of the company.

Katsoudas: I think the other thing that I would add is that everyone wants their strategies to go faster and to have greater impact. And it’s when we work together that we have the ability to do that, because we’re addressing multiple layers of that corporate stack, if you will. But I think there’s something that people wouldn’t know, which is the partnership then has also allowed on the people perspective an influence of the technology. And Jeetu, you reminded me a little bit, when we were in Covid, we were able to take insights from our people to better develop the collaboration tools that were needed in that moment and to bring some of those people insights into the technology as well.

Fuller: We’re speaking in early June of 2025, and it’s been a slow week so far for me, because I’ve only had three reporters call me to ask about something about generative AI and the future of work. So how are you experiencing that in the day-to-day work, both for yourselves and your colleagues, within the four walls of Cisco?

Katsoudas: We see it probably at three different levels. So the first is how AI intersects with our technology in service of our customers. The second is how AI intersects with the way in which we just run the company, if you will. And we see so many opportunities there as well. And then the third is, how AI intersects with your people and their readiness to do the first and second thing that I mentioned. I spend a lot of my time on that second and third area, so how the company uses it, and then how we really prepare our people.

Patel: Yeah. And one of the things that’s really interesting right now is we are moving from a world where you have these chatbots that could intelligently answer questions for us to now having agents that can conduct tasks and jobs fully autonomously. And as we have this transition happen in the industry, one of the things that’s really important culturally that has to be put in place is a willingness for people to take risks and to actually adopt these new technologies in the right way. It not only allows you to have people ready for the transition; but that transition, when done in the right way, allows us to solve problems that we would historically not have been able to tackle and solve, because you would have not the benefit of the partnership between a human and AI. And I think the combination of those actually unlock a whole new set of possibilities, that original insights will be created that didn’t exist in the corpus of human knowledge that’ll allow us to solve problems that we could have never dreamt of solving before. And that, I think, is a pretty important dynamic that requires constant experimentation and not waiting until the technology is perfected, which requires a cultural shift to be okay with being imperfect and taking risks, which I think, if we don’t do as an organization and as a country, then I think we get left behind. So I feel like this is an area where the partnership is really important because the shift is so seismic that, yes, there’s an excitement about the possibility, but there’s also a lot of fear that can be had. And we have to make sure that that environment gets to be safe for people.

Fuller: It’s interesting because what we’re observing more broadly in the economy is quite a lot of caution in large companies. Some are actually reducing their number of experiments. Many report moving from experiment to implementation, and only a rate of 20 percent to 30 percent of their pilots. And in a recurring quarterly piece of data, we’re releasing citizens in the United States are 50 percent more likely to be using AI in their personal time than at work. How are you ingraining this in the organization? What are you doing to expose people to it, to build their skills and confidence in their peers’ and supervisors’ confidence that they can use this powerful tool in a way that’ll actually be productive?

Katsoudas: We’ve recognized that the more that our people are able to experiment and use the technology, a couple of things are going to happen. So the first is, we’re going to see that their fear of the technology is going to go down significantly. Over a year ago, we released something called “Teaming with AI,” where we trained individuals within an organization on AI up to a green belt certification. And then they took that learning back to their team, and they facilitated a team discussion around how they could use a technology. What we saw was, before that pilot, there was a 62 percent comfort level with AI. And what we saw after that was that the number was in the low 90s. What we see at Cisco is, in addition to fear coming down, our people told us that they were 72 percent more creative with AI, which I think is amazing. And we are leaning into how we experiment more. We also believe that we can prove that you can use AI to actually make the workplace more human-centric, because there’s a capability that you now have to customize to your people, which I think is really powerful as well.

Fuller: Well, Fran, was that training something you developed internally or mixing and matching with external programs?

Katsoudas: It was a little bit of mixing and matching. So we basically curated a set of offerings for our people. And there were multiple levels that they could go through. The majority went all the way up to the highest certification, which I think is amazing. And then we basically captured all of the ideas from those teams. And we now have a library of prompts. We have a library of use cases that we’re prioritizing across the company as well.

Patel: The other thing that’s really interesting is, you have to make sure that you have a little bit of irrationality injected into the mix and defining what good looks like needs to have a very different kind of lens that you look at. So, for example, we have 27,000 engineers, people in engineering in this company. We spend over close to $6.5 billion on R&D. And despite having that much spent on R&D, if there’s one common sentiment that’s there across the board is, we have far more ideas than the resources to prosecute them, right? And if we can take a leadership role in saying, “What does the next 12 months look like in a highly successful state”—not just for the company, but for every individual and their career that would experience the most amount of success within Cisco—are the ones who are going to be the most dexterous in the use of AI in any job that they do. But let’s just take engineering as a job, in coding. What we then would do is say, okay, so if you assume that there’s going to be a certain percentage of code that can get autonomously written, we have to encourage people to make sure that that happens. And so, one of the things we did was equipped every engineer with tools like Windsurf and Cursor and GitHub Copilot. But we also then equipped every engineer with a charter that said, you have to figure out a way to have a certain percentage of what you do—that either you would not be good at doing or is something that you don’t have time to get done—done through AI, agentic AI. And so we actually became what the first design partner was OpenAI and Codex. And we just announced that, in May of 2025, that we’re going to make sure that we drive a level of deep partnership, even when the technology is not perfected so that we could figure out, is there a dream scenario where if we could get 50 percent of our code, 60 percent of our code, autonomously written? And that, I think, is a mental model of not just looking at the yardstick of success for the company, but also for every individual, where every individual knows that they would succeed more at Cisco if they were far more dexterous with AI. If they don’t actually have the use of AI and if they believe that AI is not going to be something, it’s a passing fad, it’s probably a wrong fit for being at Cisco.

Fuller: Fran, I’m interested in hearing your thoughts about how this might affect the workforce of the future, and particularly things like hiring patterns. In some of our research here and some of my research with my friend, Matt Sigelman, at Burning Glass Institute, we’re seeing patterns where the below rungs on the ladder may be either truncated or removed. And historically, a lot of skills have been developed basically through experience, being on the job. But if you suddenly don’t need so many entry-level people for jobs that largely now can be allocated to agentic AI, you’ve got a narrower base to your pyramid. How are you thinking about that?

Katsoudas: So the first thing that I would say is when I joined Cisco, I joined in our contact center. And so I was answering like 85 calls a day from customers. And that role doesn’t exist at Cisco anymore, because AI is taking that first line of questions for our customers, and it’s very effective. And so, Joe, something that I do see is that what we consider an entry-level role is moving up the stack. And I think that’s real. What I also have been seeing over the last five years is that companies have been working harder and harder to really onboard talent. And when we went through some of the talent shortages over the years, what we would do is, we would pivot away from experience to skills and capabilities. And we would basically say, “Hey, it doesn’t matter that you haven’t done this before. Join us. You’re going to be part of a one-year onboarding program, where we’re going to give you access to XYZ and ensure that you’re ready.” When we would do that, people would come out of that onboarding period, whether it was a long period or a shorter period, with the capabilities that they needed. I think companies are going to have to work really hard to prepare people for that next level of role, and you’re going to see more of a bridge from talent coming into the workforce and their first role. And I do think these things take a little bit of time. And so I think we’re going to benefit from those first moves. So if I use the contact center example, you’re going to have someone maybe now come into the role, where they’re the escalation level. And how do they navigate that? We’re going to have to teach them a little bit of what no longer exists from a role perspective. I do think we have to be prepared that you’re going to see roles start to come together and consolidate as well. And so from a hiring perspective, that motion also makes you focus more on skills and capabilities, because you’re not going to have many people that have done all of the roles that are coming together.

Patel: I will say this one thing: I completely don’t subscribe to this thinking in the industry, which is, the first roles that’ll go are the entry-level roles. I think it would be a strategic error to say entry-level jobs are going to be taken by AI; we’re not going to hire entry-level people. What has to happen is you have to reconfigure the job itself and say, now that you can get some of the pieces done with AI, what would you actually have humans do? And that will actually allow us to think more broadly. And to Fran’s point, that might mean that you would have broader jobs, rather than narrower jobs, when you actually get started as you’re thinking through this.

Katsoudas: Jeetu, I guess I would just say that I think what happens is, there’s a new definition of entry-level roles. And if you think about a staircase, people may not start on the first step; they may start on the third, right? And we create the ramp now to help them get there, basically. But we all kind of move up the stack a bit.

Fuller: The changes you’re describing would seem to me to make it likely that we’re going to place greater emphasis on the broad category of skills called “social skills:—the ability to interact with others in unfamiliar situations, superior spontaneous communications. It’s really, as you’ll probably know, Fran, kind of a grab bag of all the other skills we don’t have a hard-skills category for. But the type of agility, resilience, ability to handle a breadth of assignments, which suggests different types of interactions with more different types of people, would seem to me to place a big premium on social skills, which are not a skill set we’re all that adept at evaluating in recruiting processes. Does that make sense to you and what are your thoughts?

Katsoudas: It does make sense. I think we have to work really hard to make what we call “social skills” accessible to all. And I do think technology can help us with this as well. It was interesting. Last week someone said to me that they felt like, as a result of AI, extroverts were better positioned than introverts. And I thought that that was just a really interesting way of seeing the world. I think we overstate, in some cases, some of the changes that are coming as a result of AI. But if you think about the perception around the top five skills—and if I just use what the World Economic Forum [WEF] has shared recently—the first skill is around analytical thinking that is going to be most important. And then agility and then social influence. I think we’re going to have to put a lot more focus on how we really train our people to leverage the insights and the analytics that we receive. One of the other skills that came up on the WEF list, which was interesting, was motivation and self-awareness. And these are things that I don’t think we spend enough time on today that we’re going to be forced to spend more time on tomorrow.

Patel: I think one of the thing that probably is really important when we assess social skills is the ability for someone to have comfort with conflict. Conflict’s a necessary condition of business, because debate brings about good ideas. It’s people making sure that they can come to a common ground when they actually start from opposite ends. And that, I think, is a dimension that we are trying very hard to make sure that we continue to get instinctive within the culture, which is, it is completely okay to debate and disagree, and it is completely okay at the point of time of debate and disagreement to fully ignore rank in an organization, because the best idea should win. And I feel like we are, I don’t know, Fran, what would you say, I think we’re probably 30 percent of the way there and we still have 70 percent of the way to go and culturally making sure that that becomes comfortable. Most people don’t want to engage in debate. That’s not something that’s natural for us.

Katsoudas: It’s a cultural pivot. And I think it’s a beautiful example of, for the level of innovation and creative thinking that we need as a company, people have to feel comfortable sharing their perspectives and challenging others. And yeah, I hope we’re 30 percent along the way, Jeetu. But it’s a significant shift for the company, and I think it’s a great example of how our worlds really come together.

Fuller: Well, one of the penultimate social skills of the ability to engage in cognitive conflict that doesn’t spill over into affective conflict—bad feelings, hurt feelings, anger. And we try to impart that lesson to our students here at Harvard Business School in some of our courses. I’m not sure how well it always takes, but we’re trying. Do you see a sufficiently robust supply chain for talent rooted in AI? Are you concerned about the amount of talent? And where should we be looking to galvanize more investment in the AI literacy of people?

Katsoudas: The demand is much bigger than the supply. And I think that’s great in some ways, because what that is going to drive is all of us as companies to spend more time truly training and helping our people get there. There’s been an interesting pivot over the last couple of years as we’ve been hiring for AI talent. I would say that, initially, the amount of screening that we had to do as a company—so going through 20,000 candidates to get to the 150 that we wanted to hire—was interesting for us because we were looking at some very deep AI capabilities and understanding. I think now what we’re seeing is that the marketplace is getting broader. The number of people that have AI on their resume or on LinkedIn has grown quite a bit.

Patel: We did a study in early ’25 where we asked a bunch of CEOs their level of enthusiasm with AI, what their enthusiasm with AI was. And 97 percent of the CEOs said that they were really excited about the possibilities of AI for their business, but only 1.7 percent of them felt prepared. Why was the preparedness so low? There were three reasons that came out. First one was they just didn’t feel like they had the right level of know-how on getting the infrastructure ready for AI that was needed, the technical infrastructure ready. The second was you have to make sure that safety and security is baked into the fabric. But the third one was the skills needed to go out and make this happen. And on the third one, we have to make sure that, in my mind, I think there’s a tremendous shortage of people with an AI mindset, and I don’t think the answer is to go find people an AI-first mindset. I think it’s to make sure that you can actually convert folks to an AI-first mindset. You have to make sure that you bring everyone along, and that’s going to be a pretty important dynamic that every company is going to have to think through. And that requires that, one, there has to be a tremendous amount of encouragement given to try new things, have enough safety in trying new things, and not tolerate people that don’t try new things.

Fuller: We did some research similarly timed, Jeetu, and there was another consideration that jumped off the page to me in our research, which was a lack of vendor support, that certainly the LLM [large language model] companies are in very early innings for them to develop what we think of as a customer success capability and that I think a lot of large companies are used to their tech vendors providing a lot of support and almost leading them by the hand of the adoptive technologies. And that just hasn’t been available yet from the big player of generative AI, least in my estimation. What’s your view?

Patel: We recently, about nine months ago, so we consolidated our product organization. So we used to have product organizations based on business networking. There was someone that ran networking, someone that ran security, someone that ran collaboration. And then we decided that we are going to pull them all together. And I was fortunate enough to be the person who was asked to go run that. And I’ll tell you this: There would’ve been zero chance for me to have gotten up to speed as fast as I did if it weren’t for AI, because we’ve got thousands of products. We are in so many different markets. It’s impossible for one human to know all of it. But what was great was, within a matter of two or three hours of research, I could now get to a level of dexterity in an area which would’ve otherwise taken me maybe five weeks. And that level of speed of establishing competency was super important. And you combine that with what we had learned as a company to make sure that we had a very, very fast speed of establishing trust among each other. And so this notion of the speed to competency and the speed to trust between two people are the two dimensions that I think get meaningfully impacted with these technologies. And I think the element of speed is going to be really important as a cultural aspect. And for those that aren’t comfortable in driving in chaos, I think this is a very difficult time for them, because there’s just no scenario where I can predict what’s going to happen five years from now, because scientific progress, if it’s actually going to get compounded a 1,000X, what you could do in 10 years, you can now do within a year. There’s no way to look that far out. And so you have to make sure that you can feel adjust, feel, adjust, and just respond and be responsive without being reactive. And that’s a core skill that I think we have to develop as a company and as a society at large.

Fuller: Well, Jeetu and Fran, you’ve really described a very dynamic changing world, both within the four walls of Cisco and more broadly. Just a final question. As two very senior leaders with a lot of responsibility for successfully managing that transition, how are you thinking about the role of leadership in this era? What kinds of skills and adaptability are leaders going to have to show to manage this transition?

Katsoudas: I think there’s a few things that I would highlight. The first is with the significance of the changes that we’re talking about, we need leaders to create a safe space for their people. And what I mean by that is, as roles come together, as new skills are developed, people have to feel safe so that they don’t hold onto what they have; that they’re willing to let go because they trust that there will be another role and a new opportunity. And I think leaders create that dynamic. The other thing that we need from leaders at this moment is to not look at talent as “theirs” and to look at talent as “ours,” because the amount of movement that we’re going to see, which I do think is really good for people, means that we have to be understanding that there’s going to be a lot more agility to how talent navigates an organization.

Patel: I’d add one more thing, is that the era of full-time managers has now come to an end. We don’t need someone who is full-time managing other people. We need people that are actually player-coaches, because the level of credibility that you have in coaching someone goes up exponentially when you’re actually also doing some portion of the job. You have to make sure that you’re participating in the problem solving and rolling up your sleeves. So I feel like a level of extreme strategic thinking coupled with extremely high willingness to get tactical and roll up your sleeves is a super important trait.

Fuller: Well, Francine Katsoudas and Jeetu Patel of Cisco, thanks so much for sharing with us your thoughts about this exciting era, this era of generative AI and how you’re going to go about creating a workforce for the future that can exploit it and build on your competitive success in the marketplace.

Patel: Thank you for having us.

Katsoudas: Thanks so much.

Fuller:. We hope you enjoy the Managing the Future of Work podcast. If you haven’t already, please subscribe and rate the show wherever you get your podcasts. You can find out more about the Managing the Future of Work Project at our website hbs.edu/managingthefutureofwork. While you’re there, sign up for our newsletter.