Joe Fuller: What does the future hold for digital freelance work, given the growth of the return-to-office movement and a rapidly softening economy? And what does generative AI’s deployment in a growing array of sophisticated tasks mean for the firms that act as intermediaries between high-skilled talent and employers? Welcome to the Managing the Future of Work podcast from Harvard Business School. I’m your host, Harvard Business School professor and non-resident senior fellow at the American Enterprise Institute, Joe Fuller. My guest today is Hayden Brown, CEO of online talent platform Upwork. Hayden first joined us on a podcast in 2020 at the onset of the Covid-19 pandemic, just months after her appointment as Upwork’s chief executive. Five years later, she’s well positioned to share insight into another arguably more sweeping disruption in the nature of work, generative artificial intelligence. We’ll talk about how Upwork is using Gen AI to power growth on both sides of its two-sided marketplace—talent and clients—through improvements in areas like matching. We’ll look at Upwork’s pool of freelance talent and the types of projects they handle, as well as how Gen AI tools provided through the platform are making them more efficient. We’ll also talk about Upwork’s post-pandemic focus on the enterprise market and what the firm’s latest research says about the growing segment of the workforce that’s choosing to freelance. Welcome back to the Managing the Future of Work podcast, Hayden.

Hayden Brown: Thanks for having me. It’s great to be here.

Fuller: Hayden, you last joined us early in the Covid pandemic, and we’re here talking in the spring of 2025. What’s different about the market post-Covid, and what are the biggest developments that that our listeners ought to be aware of?

Brown: At Upwork, our business is all about connecting businesses, small through extremely large—you know the Fortune 100—to the most in-demand, highly skilled, independent freelancers and boutique agencies who can solve pressing business problems. Over the last 20 years we’ve been in the business of helping businesses continually transform, whether it was bringing businesses online, whether it was bringing them the services to build mobile apps, social media platforms. Based on the vantage point we have into this world of work and this ecosystem, there are a few hallmarks of the 2025 market that that I would highlight. The first one is, companies are really suffering from an AI value gap, right now. We see kind of a trough in this AI hype cycle, where so many companies are disillusioned around their ability to actually capture the productivity gains or the consumerization that they were hoping to have of this technology. And that’s really resulting in many of these businesses coming onto Upwork and looking for the highly skilled AI-equipped talent that can close those gaps for them. This is why we saw our AI categories grow 60 percent year over year last year, driven with subcategories like prompt engineering up 90 percent or more in Q4. So there’s just a lot of appetite to find this talent that companies typically don’t have in house. So we’ll see someone like a company that’s a law firm in Florida come to us and actually have a freelancer build a custom GPT for them that actually brings all of their documents and insights online and into a GPT module that they can really access. We’ll also see companies like one of the largest social media companies in the world using freelancers from Upwork to actually qualify and curate the responses coming out of their LLMs [Large Language Models]. The second thing I’d say is, resources are more constrained than ever, and that’s across pretty much every business, every industry. Companies are feeling the pinch from compounding years of high interest rates, high inflation. And I think this is why we continue to see more and more companies turning to new talent models like freelancers to help fill the gap, where they’re really needing this talent to come in cost effectively and quickly and help them do more with less resources.

Fuller: So, Hayden, tell us a little bit about your freelancers, about the population, because there would seem to be a paradox in what you’re saying. Companies really want this capability. They don’t have it internally. But rather than hiring it, they are relying on freelancers. Is that because people who are good at this prefer freelancing? Are these the apocryphal full-time employees at a tech company who are doing this on the side? What’s the makeup?

Brown: We are really playing in the space of highly skilled knowledge work. So, as an example, more than a third of the freelancers in our ecosystem have postgraduate degrees—master’s or PhDs in their field. They typically are doing complex, long-term projects. They might do a test project or something small to start, but then they very quickly graduate to long-term engagements with talent that they might be coming back to employ or to contract with time and time again. The mix of work is also quite broad. So we have 125 categories of work. Think of that as admin and executive professional support, or web development, or marketing and creative. So we have many, many categories of this knowledge work and more than 10,000 skills represented across the talent that we are serving and giving access to, which are in 180 countries globally. You can come in with a very esoteric request and suddenly find someone. This talent is uniquely skilled at using these AI tools. Because these are entrepreneurial freelancers, they are consistently shapeshifting and self-skilling, upskilling into whatever the latest and greatest technology and opportunities are, because that’s what puts food on their table. So these companies are discovering their internal teams typically don’t have the skills, don’t have the expertise to use AI tools, whether it’s in technology areas or marketing. This is actually a great alternative for them from undertaking costly upskilling programs for their employees or trying to go through those types of things, which are both expensive and time consuming, and actually have uncertainty in terms of the outcome. The other big reason we see this is, this is such an environment of major uncertainty and companies are, therefore, unwilling to put dollars behind long-term bigger investments, whether it’s hiring full-time employee pools or whether it’s green-lighting new programs.

Fuller: Well, every two-sided marketplace has got two sets of customers. So tell us, what do the freelancers look for from you? If those businesses can get access to this ready-now talent available on demand, and I can scale my consumption to what I can afford, what’s the equivalent list for freelance talent?

Brown: You know, the biggest drivers for freelance talent are work opportunities. Typically, their other options are really advertising their local market or working their networks. And these can be very limited in terms of producing the amount of visibility for their skills, which many companies might be looking for, but may not be inside their existing networks. So their No. 1 motivation is finding jobs. The second thing, I would say, is they’re really excited to be using more modern progressive AI digital tools, both to run kind of the back office of their business. So we’ve forever served these customers with, you know, automatic payments, visibility into earnings, a lot of things that they just don’t want to be dealing with through a disparate set of tools. And they also want to be delivering the work with the AI enablement that drives productivity and earnings for them. So, for example, a freelance in our platform, Marcus Grimm, is a marketing automation expert. He has actually seen his client base double, his earnings increased by more than 20 percent, and he has not been working any additional hours to achieve that because he’s using AI automation tools to do this work, both the ones provided by Upwork and the ones that he’s using kind of in his portfolio. So these are the types of things that freelancers are tapping into to basically get paid more for their time using, you know, AI enablement, which we’re helping them connect to.

Fuller: Say more about how you’re using AI in Upwork both to improve your offers and as part of your offers.

Brown: We have made a ton of progress on this very rapidly. We were, I think, one of the early companies to really commit to an AI roadmap, and that led us to do a couple of things. One is, reimagine our entire product experience as “Gen AI first.” We actually did a blue-sky exercise soon after ChatGPT launched, where we kind of thought about all of the attributes of what Upwork would need to be if we were building it today from scratch, using the most contemporary tools. And that has really delivered the foundation for our AI roadmap, which has a couple of things that customers can see and benefit from. One is, we launched Uma, an AI agent or an AI work companion that can do more and more of the tasks around, searching for talent, posting jobs, evaluating applicable talent for opportunities, more and more things in that whole pipeline to make it easier for our clients to do work using these tools. And the capabilities of Uma are just going to keep increasing as we continue to ship new pieces of that solution. The other thing has obviously been enabling the talent on our platform with the right tools, both on the like the Uma tools, which help them draft proposals, think about what rates to charge. Uma now proposes smart milestones on projects. So there’s more and more that they’re benefiting from, as well as partnerships that we’ve done to bring AI tools into the mix for them readily. We did a partnership with OpenAI very early on in this whole new era of work, and customers are using those tools, actually, to deliver the work product, in addition to kind of the work orchestration piece.

Fuller: Tell me you have something called your “agency model.” What’s that all about?

Brown: Since our founding Upwork has been home to both solopreneurs and also boutique agencies which can be, for example, individuals who start out, maybe they start building a great reputation, a great client roster on Upwork, and then they realize they could do more and expand their business and hit even bigger goals for themselves financially if they were to build an agency and basically bring on subcontractors or other freelancers from the ecosystem, or people from their own network to help expand the work that they can contribute.

Fuller: So people will essentially form semi-permanent teams made up of people who started out as sole practitioners on the platform?

Brown: That’s right. That’s one very popular agency model. We also see agency models where existing offline agencies have figured out that Upwork is a great way for them to generate new leads, connect with new clients, manage their back office.

Fuller: You mentioned that you were offering tools to the talent on the network. Do you also offer things like training for them, whether it’s basic business management training or certification, tech certs, or other steps that kind of enhance their marketability and the richness of the population on the platform?

Brown: We have a very vibrant Upwork community, where there’s a lot of peer-to-peer resource sharing. There are an array of things available for talent. But the talent continuously shapeshifts to where those skills are needed and where the demand is needed without us typically lifting a finger. They’re super savvy, and so when we actually look at some of the data around their skills, they’re at the best in the industry. So, for example, on problem solving, we have done a study recently that showed corporate FTEs [full-time employees] typically rate themselves as great at problem solving at a rate of about 44 percent. Freelancers at large say they’re good at this at a rate of 49 percent. Upwork freelancers, however, 71 percent say they’re really great at problem solving. So these kinds of differences on problem solving, on critical thinking, on clear communication come through because the Upwork talent base is intrinsically these self-starting entrepreneurs who have developed these skills.

Fuller: When you think about the evolution of AI just within the firm, and how it’s brought more dynamism to your platform, what are you learning about things like algorithmic management and monitoring the performance of your algorithms and the how you how you measuring the performance of your experiments, because if you look at the aggregate data—and I think this is borne out by your description of the way a lot of your customers are seeing the market right now—there’s cautious experimentation, a lot of companies with generative A,I and the success rates that are being reported are pretty unimpressive.

Brown: It’s true. And our experience here is definitely anchored in the fact that we, for almost 20 years now, maybe more, have had a work diary product, where talent can log their time, and that automatically turns into billings. One of the things that I’ve heard throughout my entire career from employees across the different businesses I’ve been at has been, typically, they’re not getting the level of feedback and visibility that they really want from existing kinds of management structures. AI-augmented management means the talent can have a lot more visibility, a lot more data, a lot more prompts into how they’re doing, how they can improve, how they could, you know, do things on their own without the manager having to come in. But at the same time no one wants to be managed by an AI. They want all of the pieces of what an actual human can do in conjunction with those kinds of data-driven insights. You know, we’re very early in that journey, and so a lot of these AI tools aren’t built out yet to really do that effectively in hand in hand with humans.

Fuller: It’s going to be really interesting to see how it all evolves, given the rate of improvement, I would argue, in the last six quarters is faster than we thought it was going to be in terms of the technology. And there’s some awfully interesting new studies coming out, for example, of AI-powered coaches being equally effective in providing feedback to workers as human coaches and significant productivity increases in R&D functions with declining job satisfaction for the knowledge workers. So there’s so much to learn about how this is all going to play out and for us to be able to manage this transition more effectively. Take a step out for a second. You get a great look at the demand side, what companies need, particularly where they have shortfalls. And you also have a good insight into the U.S. talent pool. Are we producing enough of the right type of talent in your view in the United States?

Brown: Let me first respond to the last comment you made around the rate of change in this area and what’s happening, because this is an area where we are using some of the most unique and large-scale data sets that we have at Upwork that really capture what type of work is happening and what are all the steps happening in those workflows and using that visibility into the full work lifecycle to train our own agent, Uma, to be able to do more and provide more real-time feedback and more real-time interventions to help projects and work stay on track and help both parties have things like ongoing, clear communication and timely updates and things like that. To your question about, are we producing enough white-collar talent with the right skills, you know, in aggregate, the data actually shows that we do have in many, many of these categories enough talent, kind of globally at large, to do this work. But the problem is businesses are typically not looking in the places where that talent exists. So there’s a real mismatch. For example, more than 50 percent of Gen Z workers are already freelancing. This is a huge stat. I mean, this new generation is the multi-hyphenate generation that wants to have a career composed of many different things in a portfolio, not just one linear, full-time opportunity. And yet, since many of these Gen Z folks are freelancing—and a lot of companies are still figuring out, “How do we incorporate freelancers or where do we go to find them?”—they’re using outdated tools like their staffing firm or LinkedIn to look for these people. They’re actually not connecting to where these individuals exist on more modern platforms, such as ours. And so that’s often where the disconnect is, you know, talent is one place, and clients and businesses are in a different place.

Fuller: One reason I appreciate Upwork is, you regularly do some of your own research, and that can form some of mine. You’ve got the Future Workforce Index, where you track some of this data. What are your most recent findings from that?

Brown: The headline here is that freelancing continues to grow as a percent of the overall workforce. Today, it’s more than one in four skilled workers who are freelancing, and they’re producing more than one and a half trillion dollars for the economy in the U.S. alone. It’s continuing to be more and more popular on the talent side because of the things around freedom, flexibility, autonomy that this talent is really looking for. It’s also super interesting that, whereas maybe 10 or 15 years ago, a lot of folks were freelancing out of necessity, or they felt like this was a job of last resort, that equation has totally flipped. Generally, what we see is, talent understands that the employment contract is broken and has been broken for many years. You know these are folks who’ve seen themselves get laid off or their parents get laid off, and the economic hardship coming from that. And so, increasingly, individuals are seeing that building a freelancing career—where, on our platform, for example, the average talent has five clients in their portfolio— they know that they can withstand economic turmoil, uncertain environments such as today’s environment by really, leaning into this way of working. You know, one example of this is Jamie Hollander, one of our top freelancers on the platform, who I first connected with during the pandemic. And I will never forget. When I got on the Zoom call with her and she said, “Hayden, I’m calling you from the house that Upwork built.” And she recounted how that, you know, journey from being a full-time employee in a media agency in New York City to then moonlighting on Upwork and then making the big switch to fully invest in our platform, working independently as a freelancer and then building an agency had just transformed her economic situation, her prospects, her feeling of optimism and stability, even in the dark days of Covid, when so much was going on in the economy. So we see in the data that this is how talent are moving, you know, their work aspirations. And we also see more and more clients, of course, waking up to the reality that this is where they need to go to get these highly skilled workers.

Fuller: What do we know about their job satisfaction in doing this work? I can imagine kind of doing tasks, particularly the same type of task again and again and again, for people you don’t really know and don’t really have a relationship might be not all that motivating to somebody. But also, there’s the freedom of choice and all the other attributes. What does your talent tell you? What does your survey of their attitudes tell you?

Brown: What we see is that they are really excited about the work on Upwork, because it actually continually keeps them fresh and exposes them to either new ways to apply their skills—there’s a new business context, a new project—or ways to actually deepen their skills, because they’re very focused on the area of expertise. So many freelancers are interested in building careers this way, because it really gives them the ability to go both deep and broad, depending on their appetite, depending on the season, depending on their interests, rather than be pigeonholed into something that maybe in a corporate work environment is less fulfilling and less educational, frankly, over time.

Fuller: Well, I can readily imagine that being able to pick the work you want to do and bid on, as opposed to do the work that your supervisor asked you to do, could be quite an uplift for a lot of people.

Brown: Yeah. And I’ll add that you know businesses win in a huge way from this as well, because what we also see is that freelancers are more engaged, they’re clearly more skilled, and they’re more equipped to dive in and add value immediately out of the gate when they’re brought into these different projects.

Fuller: So do you have a theory of the case, Hayden, about how this all plays out, and what type of mix of employment circumstances we’re going to have in 2030 or 2035?

Brown: I think I would be pretty foolish to make a prediction that far out, Joe, given this rapid pace of change that we’re in. But I will say I am an optimist. I understand that this change is really scary for people. I hear that from, you know, individuals across the economy all the time, saying, “Look, this might come for my job,” or “I’m not sure I can learn these new skills,” you know. Typically, that’s not what we hear on our platform, but certainly from people in the broader economy. Those concerns are very real. And yet, when I look at the data, I am really heartened, because what we see on Upwork is, AI augmentation is actually driving increases in jobs; it’s driving huge increases in earnings. So the average freelancer that’s using AI skills and tools is earning 44 percent more than someone not using these capabilities. And so, while I do think we will see some disruption and dislocation from changes, people can cross the chasm. I believe they are agile, they are resilient, and they can actually learn these new skills and tools. And it actually creates more opportunity for them. I also see the data that says companies that are using AI at the highest rates want to hire more people, whether freelance or full-time, as they’re getting more productive and actually unlocking so much of this with their businesses. We’ll see a shift toward more freelance and independent work styles, because it is really a win-win for everyone, even if the core size of FTE pools in many corporations continues to shrink. That doesn’t mean that the total opportunity is getting smaller. I think the opportunity, actually, will continue to grow for people working with these tools.

Fuller: Certainly, an exciting consequence of having more platforms like Upwork accessible to workers is it also removes barriers to workforce participation by people who are in some way marginalized. We’ve certainly seen that in our research—that they are a large quantity of what we call “hidden workers,” who are either on the fringes of full-time labor or working part-time, anxious to raise their income levels, but find it very difficult to make the transition to traditional full-time work out of that part-time platform. And the growth of the gig economy, whether it’s for high-skilled white-collar labor like you provide a lot of your clients or more traditional two-sided marketplaces that we’re familiar with, like Uber, it definitely provides alternative income streams that can help bridge economics for workers in transition.

Brown: You’re absolutely right, Joe, and this is really a key through line for us. When I look back personally at how I grew up, you know, I was inspired by the role model of my parents, who both worked in international development. We lived overseas in Kathmandu, Nepal, for many years as a kid, and I saw firsthand that talent everywhere was just incredibly skilled. And yet the economic models and the labor models that were predominant at that time really required people to move to cities or move to countries or continents to get to where those work opportunities were. And yet today, fast-forward to where we are at Upwork. And why I joined the company 14 years ago is really because of this true meritocracy that is online, that’s digital, accessible by everyone, both businesses and talent, to get to these opportunities that don’t require people to relocate and disrupt their personal lives and everything else. So it’s really exciting that we have a for-profit model that really supports so much advancements around economic opportunity, at the same time as recognizing that the old models of work are really outdated.

Fuller: And I can’t let that story go by without asking you to tell me a little bit about what it was like to grow up in Kathmandu.

Brown: You know, it was a special experience. I think, because my parents did all this work around microfinance and literacy, and they were really invested in the local community, I got to see firsthand, you know, what people were struggling with, where people were succeeding, where the economic system was not serving them, where political institutions weren’t serving them. And that was really the motivation for me to do so much of what I do today. And so it really came home to me that opportunity in the world is really not evenly distributed, and we are so fortunate to be, you know, in the U.S. and have all the benefits and privileges of what that confers on us. So it was an amazing experience. It taught me a lot, and I think it’s one of the reasons I’m here at Upwork today.

Fuller: So let’s talk about your recent past and then what we can expect from Upwork going forward. In ’24, you did a pretty big restructure, it looks like, and a redirection of the company. What motivated that, and how is it positioning you for the future?

Brown: So last year the job to be done for us was around continuing to increase our profitability profile, because that was a key way for us to control our destiny as a public company, while also driving a more limited portfolio of growth investments. So we did a lot of deep work, pruning different things in the business, realigning resources, and frankly embracing the start-up roots that I had experienced in multiple points in my career, including my early days at Upwork when we were a tiny start-up. And so, coming out of this, we are moving faster than ever. We have the agility of a start-up but the depth and breadth of a market leader in our space, which is really what we wanted to get to, to attack today’s moment.

Fuller: So, with that repositioning, what can we look for out of you in the next two, three, five years.

Brown: Yeah, the next few years are, I think, going to be the most exciting ones that this business has seen. We are really leaning in on the AI reinvention of work, putting together the best of human intelligence with artificial intelligence to do more for both talent and businesses around the world. We’re also continuing to move our company upmarket, in the sense of serving larger and larger customers through our enterprise strategies. There’s many companies that have more opportunities to tap into these incredible talented people and the automation and the acceleration we can provide with our AI tools around orchestrating that work. So those are going to be two big vectors for us—the AI enablement and the enterprise enablement, specifically, as we continue to expand our offerings.

Fuller: Well, Hayden Brown, CEO of talent platform Upwork, thanks for joining us again on the Managing the Future of Work podcast. You’re one of our few two-time guests.

Brown: Joe, it was such a pleasure. Thank you for having me.

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.