I think with that, I know there's going to be people that. It's amazing how in five minutes people drift in and out of rooms. But let's get started in the interest of time, just to keep the trains running. It's my pleasure to have Upwork back at the conference again this year. Hayden Brown, CEO. Hayden, you're always so generous for coming every year. Thanks for allowing us to have the conversation.
Thanks for having me. Absolutely.
Let's just start. I always do like to start a little bit big picture for the folks who don't know the company as well. Talk a little bit about what you're building, what you're scaling, what your key strategic priorities are, and then we'll build off that answer and go a little bit deeper and everything.
Yeah. Upwork, I think has always been more than meets the eye. I've been at the company for 15 years, and I think one of the things people discover when they really get to know the business is it's much more than just a talent matching platform or a place for people to come find freelance talent. We really serve customers on both sides, clients, which are small businesses through large enterprise, as well as independent talent, with the full end-to-end product around not just finding great talent or finding great work, but also delivering on and getting paid for or paying for that work. It's really an end-to-end platform that spans the entire work life cycle, and so we call it a work marketplace because it's more than just talent matching.
I think that's particularly important in the era of AI, when so much is changing, talent discovery methods, recruiting, all of this is, I think, being rewritten in real time. For us, our ability to provide a trusted platform that gives clients and talent the knowledge that what they need to get done is going to get done, and that they'll only pay for work that's been delivered is incredibly important. I'd say in the last 12 to 24 months, some of the thematics for us have really been around three growth levers that are really working in the business right now. The first of those is our push and our expansion upmarket to serve larger SMB customers. What we're doing there is a custom product called Business Plus, which is really meant for larger businesses operating on our marketplace.
It gives them a set of tools, including AI recruitment tools, including access to net 30 payment terms, access to collaborative hiring features, all of the things they need to really run their business at larger scale and with larger programs using independent talent on our website. That product is really our fastest-growing in company history. It grew 174% year- over- year last quarter and is ahead of plan. We had set a goal for it to be 5% of our total GSV this year, and already it's at 6.6%. I t's growing super well. We're seeing a lot of demand from customers that are bigger, looking for this type of offering. The second growth strategy we have is also around moving upmarket, but this time into large enterprise. In this instance, we've had an enterprise business at Upwork for many years.
Last year, we actually reconceived and relaunched that business, and really had rebuilt it end to end. We bought two companies that now are part of that new Lifted business that give us broader ways to serve large enterprise customers. This is a huge market. It's $650 billion of enterprise spend. Now with Lifted, we're able to serve all of that spend inside customer accounts across not just independent talent that's doing 1099 or independent contractor arrangements, but also other engagements as well. W e can do staff augmentation. We can do EOR, or employer of record, AOR, agent of record. We can do managed services and SOW work. That change to our Lifted strategy means that now we can sell into the largest enterprises in the world, a solution that really meets the needs of their compliant contracting and contingent workforce programs.
We can sell into those programs in a centralized way to compete for much more of those budgets, which typically are in a single account, might be $100 million or more of contingent spend. The Lifted strategy is off to the races. We're seeing really great traction and pickup with our sales team and with conversion in that business. That's something that's also performing really well for us this year and is on track for 25% GSV growth in 2026. The third piece of our growth strategy is really around AI. I'm sure we're going to talk a lot about AI today in this discussion because it's a big thematic for us and certainly is, for us, a tale of both headwinds and tailwinds.
From a tailwinds perspective, what we're seeing is really the acceleration of demand for talent on our platform who can help small businesses and enterprises with their AI-related needs. This entails all of the work around things like AI video generation or AI strategy consulting, which is a category that grew more than 50% year-over-year last quarter. We're seeing lots of demand for this type of talent because SMBs are trying to figure out, how do I adopt AI? Where do I get the talent that can do this for me and get value out of this new technology? The answer is they can find that talent right on our website.
The other piece of our AI thematic is really around our own tooling and rebuilding of our platform to be AI native, specifically to actually integrate into all of the AI tools such as Claude or ChatGPT, which customers are using to get work done. Just in August, we launched our MCP, which means now customers inside of these AI environments can invoke Upwork and get talent directly inside those environments from our platform, and can also execute their workflows related to Upwork from within those tools.
This is also the stepping stone or the building block for us to expand our marketplace to not just serve human talent and human business owners, but also their agents, and the agents that are coming to Upwork to actually find human talent, which is a trend we've already seen happening, as well as enabling the ability to bring agents into the equation to deliver work on behalf of clients over time. These are really important growth building blocks as we think about where the business is going and evolving it not just from a demand standpoint and a channel standpoint, but also in terms of the core offering and who can participate.
Okay. There's a lot to unpack there, and I think we're going to go down most of those roads as we talk, but maybe just to level set with what you've learned about AI over the last 12 to 18 months, you highlighted some of the headwinds and tailwinds that the business is facing today. How has it informed any changed view about the long-term opportunity that presents for your company as a result of AI?
There's probably two things. One is, this is a business. We are a broad horizontal platform that serves more than 130 categories of work, more than 10,000 skills. Basically, every type of knowledge work happens on our platform already. That means that through every past era of work changing, whether it was the advent of social media or the introduction of customers trying to move things to the cloud, every time work has changed, we have seen those changes on our platform, and we've supported the birth and death of every type of work that's happened in the last 20+ years. AI is in many ways no different. There's new demand that's forming around expertise and skills that people didn't even think about a few years ago, and now we're seeing those across our platform, and we're home to more than 300,000 AI practitioners globally on Upwork.
We're also seeing the death of certain types of work. Again, that story is not new for us. I think the only thing that's different now is the speed of that happening. The AI categories are growing faster than any categories we've seen in our history. We're also seeing, at the low end of our market, automation and displacement of work at a very rapid rate, which is, again, a cycle we've seen before, but it's being compressed right now. I'd say the other AI thematic for us that's really emerged over the last 12 to 24 months is this idea of moving Upwork from being a destination to being infrastructure. What I mean by that is previously, in the pre-AI era, you had to come to our website and sign up and manually take steps and actions as a customer to get the value of Upwork.
Now today, with our MCP, you can actually get all of that value through directing your agent to act on your behalf or bringing forward into your own third-party AI tools the value and the talent that Upwork has to offer right inside those environments. This is an important evolution for us as we think about the world moving, in terms of work moving into these AI tools and more of the demand sitting there with customers that are executing work inside these tools.
Now Upwork is integrating directly into those experiences and can be called upon, including with true on-demand talent solutions, where you could invoke a fractional CFO or a marketing leader immediately from our platform to come in and review and participate in work that you're doing inside of a tool like Claude. This is really a new direction for us as we expand how and where we're serving customers given AI.
Okay. Let's turn to Business Plus. What do you believe is resonating the most with customers with respect to Business Plus, and how do you think about the market opportunity it continues to open up broadly against the SMB landscape?
Yeah. The number one thing that customers tell us they love about Business Plus is the access they get to our highest quality, most curated talent in the ecosystem. They get preferential access to that talent. They also love the AI recruiting tools that do things like AI interviews on their behalf. They can ask our AI agent, Uma, to review and generate a shortlist of candidates from an applicant pool, to actually go and interview those candidates and bring back snippets and a summary of who is a good fit based on that specific client, that specific job. These are some of the things I think are really drawing in customers on the Business Plus side. Furthermore, the ability, Eric, to scale their work across multiple participants.
We have a multiplayer mode, we call it, in Business Plus, where they can really collaborate around work, around hiring, have multiple participants. Again, if you're a larger small business with a team or multiple departments that are using Upwork, these are the types of features that they really need. This is proving to be a great, not just upsell for existing customers who are in our marketplace being very active, but also an attraction and an entry point for new customers who are coming in. We see them on Business Plus spending 2.5x as much in Business Plus as they do coming into our standard Marketplace Basic product. We also benefit because this is a higher take rate product. We monetize this at 10% on the client side instead of 5%, which is how we monetize our basic product.
As we're seeing this really fast growth in Business Plus, it's helping us not just drive GSV and revenue from these customers, but also our take rate is improving as well.
Okay. You mentioned Lifted earlier. Can you talk a little bit about how that is also changing the nature of the business in terms of opening up the larger enterprise opportunity?
Yeah. This is such an exciting one for us because the market is so huge, and the customers we've been serving enterprise for many years have said to us again and again, "Upwork, we love your talent. The talent we get from you is better, it's more qualified, it's more engaged, everything about it, and frankly, it's more cost-effective because it's across 180+ countries globally." But they've said, "We don't want to just consume it as independent freelancers. We have these other programs that by definition need to be serviced with different contract types. Can we please consume your talent in these other ways?" Historically, the answer was no, because we had not built the in-house capabilities to do that.
But since we launched Lifted last year, now we can say yes, and we are winning RFPs and are in contracting with customers for these new engagement types that include staff augmentation, employer of record engagements, things like that, which previously we were not able to access. This has really changed the conversation with our customers, both the existing customers who know us in certain areas and have wanted to expand with us in other ways. It's also letting us land new customers who would never have historically considered working with us. Just last quarter, we landed one massive multinational company that spends probably north of $100 million a year on contingent work. They are now contracting with us for a type of work that we were not even eligible to compete for in the past before Lifted.
I think we're really seeing with our sales funnel and the feedback that this is a very differentiated offering, and it's something that's really opening up a lot of new doors with both current and new customers.
Okay. One of the questions I get the most is as you go through this transition to where AI is taking the business longer term, how do you think about which clients to emphasize from a growth standpoint versus elements of clients that are in transition, where maybe the nature of the work or the nature of the relationship they have with the company won't be where you want to be from a mix standpoint a couple of years down the road, and how you think about managing through that transition?
Yeah. We've been very focused on a shift that we've really executed over the last couple of years from more of volume-based customer acquisition and servicing to much more of a value-based approach. That shows up both in our marketing spend and our acquisition, our digital acquisition, where we're focused on targeting and converting these larger SMBs who can self-serve in the marketplace, and then, of course, through our sales channel, larger enterprises on Lifted. It also means that we're building out and putting our product dollars against these newer products like Business Plus, which we know can really attract and convert a set of customers that were maybe bypassing us or not thinking about us before because we didn't have these custom offerings for them. I would say for us, it's a view of expansion.
We're not saying goodbye to the smaller customers, the very small businesses who are still the bread and butter in our core marketplace, but we know that there's a lot of smaller transactional work that a lot of these customers do that doesn't represent the future where the business is going. We're really building both for that future that we can already see in our numbers around larger customers with more complex work needs, spending more on a per-project basis. We hit, this past quarter, another record high in terms of GSV per client, which is over $5,000.
We've seen continued increases in spend per new client because these acquisition approaches are bringing in larger customers, and that is going to continue to happen, I think, as we grow with these bigger customers and just focus less on some of the smaller transactional stuff that even may be here today, but over time, will be gone with AI in the future.
Okay. To that point, what do you continue to learn from the most forward-leaning clients with respect to AI as to what AI native demand is going to look like for this platform somewhere down the road when there's a greater percentage of the mix being driven by AI first companies?
Yeah, we have some real enthusiasts in our customer base who are on the bleeding edge of this stuff. I think there was a moment earlier this year when a customer was saying to me, "I really wish Upwork would let me use all of what you do inside these environments and with these use cases where I could super scale my work on Upwork because I could have AI 10X-ing how I use the platform today." That was a great moment because that was exactly what we were building with the MCP, but hearing it from customers who are saying, "Look, give me more access to do more on your platform through this capability," that's kind of a dream come true.
I'd say we're seeing on the talent side, these very pioneering customers who want to scale their activity, they want to use AI to do more and get more leverage themselves out of their time and their serving clients on our platform. On the client side, we see businesses that want to be integrating more of this talent into more of their workflows, but they want it to be seamless, and they want it to be inside the AI environments that they've already invested in building out and tailoring. This is really where we're taking the product, on behalf of these customers and where they're showing us they want to go.
Okay. One of the other things we have talked about on some of the more recent earnings call is agentic workflows and how there is going to be agents that talk to other agents.
Yeah.
You sending your agent out into the world. How do you think about positioning the platform medium to long term for the nature of work and sourcing of work becoming more agentic workflow dominated?
Our view is work is going to continue to, frankly, expand, and there is always going to be a portion of work that is purely human delivered, and it needs that human touch. Then there is going to be a portion of work that clearly machines are best at, and they can automate, and they can do that work autonomously. Then there is going to be this big, messy middle, where there is a lot of work that maybe AI can be part of, but humans need to supervise or qualify or approve. We are basically extending our platform so we can serve all three of those work types.
Okay.
We definitely have always been the leader on the human-to-human work and the marketplace we have there. Now we have humans and AI agents working collaboratively together on our platform using tools like our MCP, and as agents get more capable, we will be able to support them with this MCP layer to actually directly deliver work on the talent side, or again, call into our system and invoke humans or agents to do work on behalf of clients that they are operating for. We view this as just a big continuum, and our goal is just to serve customers across all of it. As work moves from one bucket to another, we are kind of indifferent because we can basically service all those types of work right there on our platform.
Okay. One of the other topics you and I have talked about before is just the evolution of discovery on the platform.
Yeah.
Talk a little bit about how you plan for the product roadmap to continue to amplify and improve upon discovery more broadly so you are getting higher quality matches, higher conversion, all those dynamics that come as an output of discovery.
Yeah. You are talking about within the platform versus- client acquisition?
Yeah, y eah.
Yeah.
We can talk about both.
Yeah.
Whichever direction you want to go.
They are both relevant. I would say, within the platform, we have this treasure trove of a data asset. We see 3 million jobs posted every year. We don't just see the jobs, we see every candidate that applies to those jobs, who gets interviewed, who gets hired, who gets shortlisted, and then the full funnel of that work being delivered. That data asset that goes all the way through the work delivery and the payments gives us an incredible set of information to train our matching algorithms against, and use that full 360 view of what happened in every single interaction to then propose the right next candidate for the job and the right next job for the right candidate. W e are tuning that engine all the time.
I would say one thing we have started to do more of as the work on our platform has continued to get bigger and more complex is build out our search and match so it is actually even higher fidelity at taking a client requirement, which may be expressed as a job post or a query, and decomposing it into the subunits of work or tasks that actually might need to be executed either by one individual or maybe by multiple different experts within our ecosystem. Then using that X-ray view of the underlying work that is needed to be done to then go into our database and do all of the matching around that. This is obviously something that we are continually iterating on.
We've made some, I'd say, leaps forward in the last couple of years with some acquisitions we've done, especially of a company called Objective AI that brought us some really great kind of AI native search capabilities, which now are fully integrated into our stack. As the work keeps evolving, we continually tune the system for those new flavors of work and the more complex work that's happening on the ecosystem.
Okay. Going back to the investor day you had towards the end of last year, you've been talking about the role that data can play against the broader AI opportunity. Maybe talk a little bit about how you think about the data you have as a company, the value of that data, and how to think about ways in which data can turn into monetization and opportunity.
Yeah. This is an interesting one, and we're still early in figuring out how we can fully deploy this data asset, because, again, what we have is not just kind of a static snapshot from a period of time. We have the full history of work on Upwork. But every day, that data asset is getting refreshed and updated with the most current jobs, the most current skills, what's in demand in the market, who are the practitioners and what are the things they bring to the table that make them really good at doing that work. There's multiple ways we've thought about applying this data, and it's really been approached by different kind of ecosystem players who are looking for either training data around specific types of work to tune their AI agents or their models.
Sometimes they're looking for actual work products as an input into how they can improve their tooling and their solutions. Of course, there's also the aspect of customers, some of them care a lot about and want to keep their information very private, and others actually could care less, and they're like, "Hey, have at it. This data doesn't matter to me." We're kind of looking at all those things, like what's the demand in the ecosystem? Where are our customers in terms of what they're comfortable having shared or not? Using that to inform, I think, what could be interesting opportunities going forward. It's still early days, but certainly the data asset we have, we keep kind of coming up against the reality that it's very differentiated.
No one else has this size of a data set around real-time, real-world work getting conducted that has a digital trail to it and that is completely visible and readable by both humans and machines. I think that is something that we will continue to explore and figure out what can we do with this one.
Okay. When you take a step back and you think about what we have talked about so far, the evolution of the platform, the things you want to invest in to make sure you are capturing some of that growth, bring it back to a conversation about how investors should think about capital being allocated? You are making sure you are making the growth investments you want to make.
Absolutely.
You are also probably going to try to deliver on the margin trajectories, you have talked about, and then also the scope for the potential for either M&A to grow the business, which you have pulled that lever before, and/or thinking about capital returns. Talk to me about the process which the management and the board goes through to think about optimizing for that mix in the business.
Yeah. I would say, first and foremost, we know this is an extremely high margin business and one that we can continue to dial the dials on in terms of margin, both because we have incredible discipline in managing our cost base. We also are growing our highest margin products the fastest of anything in our portfolio. We see ongoing opportunities to kind of optimize where and how we are deploying capital. Margin expansion is certainly something that we continue to foresee for the business. I think it is something that we have demonstrated we can do and will continue to do. We also want to be sure we are investing for these incredible growth opportunities that are unique to our business.
When we went through our 24% reduction in force earlier this year, we were very mindful of actually concentrating more resources on the things that are really working, our AI strategies, our SMB build-out, our Lifted build-out, and investing in those things and bolstering them further, even as we streamlined and stripped away either other things that we had previously been investing in that we thought were no longer relevant, or did things like optimize spans and layers in the organization and kind of how we are managing the business very efficiently. That is going to continue to be something that we do. As this business continues to generate a lot of capital and cash returns, we are basically executing a capital allocation strategy that spans share buybacks, on the one hand. We have done about $110 million of share buybacks this year.
We have $254 million left on our authorization from the board, and so that is certainly going to continue to be a priority. We will always reserve a good amount of capital for operations and some dry powder for M&A, because certainly in our space, we are constantly looking at ways that we can accelerate our growth strategy with M&A. We do not need to do M&A to achieve our goals, but certainly there are opportunities that can come along that are very attractive. We are really, I think, able to balance with our strong balance sheet, investments across all of those buckets and continuing to improve our margin in the backdrop of that.
Okay. Then maybe last one just to wrap up. When you think about what you are most excited about, and then you measure that against the debates and the conversations you have with investors, where do you think there is a gap between what is underappreciated today about where the platform is going, especially when measured against what you are most excited about delivering against over the next three-five years?
There are probably two things that I think are very underappreciated about the business today. One of them is its durability, and both the fundamental profitability of the business model and the durable moats that we have built through investing in a two-sided marketplace as the pioneer in our space that competitors have come at time and again and really not made a dent in our business. I think the durability of this platform, I think there is an over-indexing of fear around AI automating work, which is certainly happening in pockets, but I do not foresee a future where humans do not have jobs. I think there is going to be a lot of work for people on our platform to do.
There's an underappreciation, I think, of the durability of the model and how well that's going to perform once we're through just this short-term transition of what's going on around both AI and also some of the acquisition-related headwinds from Google and other things that are impacting the entire industry. I think the other thing that's underappreciated is just how exciting these opportunities around becoming an AI infrastructure play can be for our business. Because no one else can do what we can do here. No one else has a talent pool of 18 million professionals that are ready to work on-demand, and we can deploy through a technology platform into the full ecosystem of work that's happening. Frankly, with AI, more and more of the work is moving online and into these technical tools versus sitting offline inside of offices.
As that work moves into these tools, we then have access to so much more of that ecosystem and sending our workers into these endpoints to do this on-demand work, which I think is, again, expectations are changing. With the advent of AI, customer expectations are that they need everything now, everything real-time. They want intelligence in real time from AI, but they don't always get what they want, and therefore, they need human solutions to augment that, but they need it in real time. That is something that, again, Upwork is uniquely positioned to do. That really sets us up, I think, with an incredible set of growth levers that we can lean into, both on the AI side and then also, of course, with these larger customers. I think time will tell and demonstrate the value there that's maybe not fully appreciated today.
Okay. Hayden, I always appreciate the opportunity to have the conversation. Thanks so much for coming to the conference. Please join me in thanking Upwork for being part of the conference this year.
Thank you, Eric. Thank you for having me.