Okay, welcome back. Hayden Brown is President and CEO of Upwork. Hayden, welcome.
Thank you.
Good to see you, and thanks for supporting the conference. Maybe just level set everyone with what's going on in the overall environment, the business, what you're seeing from your perspective at this point.
Yeah. It's definitely been a very interesting year so far, and in many ways not the year we expected, but the business is doing great, and we can talk more about some of these trends. I would start by saying, in Q1, we saw some emergence of headwinds, specifically related predominantly to macro. Just so I can unpack that a little bit more, we look at our business, we've got about $100 million of enterprise revenue. That was unimpacted. That was completely unchanged. That part of the business is on track. We're expecting 25% GSV growth in that area this year. Those customers were really not seeing the pain. Also, we have an SMB product called Business Plus. We launched it about 18 months ago, and customers in that product also did exceptionally well in the quarter. We're ahead of plan there.
That product grew 34% GSV quarter-on-quarter. Where we did see this kind of dynamic environment was actually in our Marketplace Basic plan, which is about 85% of our revenue. The average customer in that plan is a very small business that has on average fewer than 10 employees. Think of these as the most exposed segment of the business side of the economy in terms of companies that really don't have a lot of cushion. They've got to make fast changes when things are moving. We saw in late February, these customers start to spend a little bit less, so their contract volumes came down, and that progressed for a number of weeks in Q1, and then leveled off in April. When we asked those customers why were they spending less, it was very clear.
They said, "The economy is impacting my customers and my business, and I just don't have the budget I used to have to spend on not just Upwork, but labor and related services in general." Interestingly, as I drove over here actually from the airport this morning, my driver said to me, "Hey, do you work at Upwork?" I said, "Yeah." He said, "I'm an Upwork customer. I run a fleet of six vehicles. I'm a small business. I've got two employees, me and my wife, plus some folks who do the driving, seven people who do driving for us." I asked him, I said, "How has business been?" Because he was like, "I use Upwork for marketing.
I use Upwork for a bunch of things for my small business." His answer was, "This has been a really tough year for us." He said, "It's been very up and down," and he was doing this hand motion while he was driving. I was like, "Don't drive off the road." He was like, "It's been very up and down, and we're seeing in our business headwinds from less consumer spending." He said, specifically, people aren't doing things like renting his buses for bachelor parties and big nights out. He said also customers that used to fly in from Canada and Australia to come to Disney are just down with jet fuel prices, et cetera. He was like a perfect snapshot of what we heard from customers, where they just don't have the cushion, and they're seeing this volatility, and they're spending less.
That was the biggest impact we saw in the quarter and the biggest driver of our top-line guidance change. A secondary effect, which I'm sure we're going to talk more about, was AI, a smaller effect there. We can talk about that now or later.
When you think about besides what the issues and the macros, anything else change in the core business in terms of everything else seems pretty-?
Everything else is stable. Again, enterprise and SMB are looking really good and are on path. It's really been with these smallest customers where we saw this volatility, the volatility was widespread in terms of categories of spend, customer types. It wasn't just in one work type, which again, is I think a clear indication it's not AI that's doing this type of work. It's more that overall spending was just under pressure for these customers, that was really the notable change in the quarter.
To have 25% GSV growth in the enterprise, that's really healthy. What's driving that?
Yeah. As you know, our enterprise plan was expecting back-weighted growth this year, so that 25% GSV growth is going to come really in Q3 and Q4, and that's due to the work we did last year. We made two significant acquisitions for our enterprise business to bring on some critical capabilities that expand how we can serve enterprises in terms of the contract types that we can support. Previously, we were phenomenally well-positioned to do independent contracting work for large enterprise, and some agency of record work. These are different kind of classification types for the workers. When it came to things like EOR or more payroll relationships, which is really where 95% of enterprise budgets for contingent work go, is they do want EOR and payroll solutions from their vendors like us. We didn't have those capabilities in-house.
We bought a company, Ascen, last year that brought that capability to us, and on the back of that, we've seen tremendous expansion in our enterprise pipeline as we've been ramping up sales this year with the platform going live in June. Our new client pipeline was up 3x in the quarter from January to March, and our existing customer pipeline was up 9x in terms of the number of engagements. That's really setting us up to be migrating customers to this new platform in June and then ramping those programs in the back half of the year, and that's what's going to drive that 25% GSV growth.
34% growth in the SMB business?
Yeah. That was quarter-on-quarter in Q1. Just to put that in context, today that is a new product. We launched it 18 months ago. It's our fastest-growing product in history. It's targeting businesses that have 250 employees and higher that are in that small to mid-market sweet spot. It only accounts for about 5% of our GSV today. That product has been ramping quickly since its launch a little over a year ago, but we need it to continue to scale to support the top line further. It will, as it does that. Again, as we're seeing these puts and takes between very small businesses that are having some more economic turbulence right now and the larger small businesses that can really ride through that and make bigger decisions, that's where we're seeing a real bright spot that will keep scaling.
Now we're onto AI.
Do we have enough time?
Internet conference. For AI, is this just these small businesses are like, "Look, I don't need to hire a Adobe creative professional. I can do it now on ChatGPT"?
What are we seeing there?
Well, that's what I'm doing, personally.
Yeah.
What are you seeing in terms of what's having an impact?
Yeah. What we've seen actually for several years is at the very bottom of our market, you think about on Upwork, the average contract size or really spend per client is about $5,000. We have everything from very small projects all the way up through hundreds of thousand dollar and larger engagements, of course, multimillion-dollar engagements in enterprise. We've seen for several years now that the smallest jobs in the marketplace are getting automated. These ones are more like discrete tasks versus bigger, complex role-based work, which has not had that same impact. We did a new analysis in Q1 as we saw changes in terms of client volumes, and we were ourselves constantly looking at how much of this is macro, is it AI, is it other things?
What we saw was Anthropic and Stanford had both released some new studies around how to measure AI job exposure in terms of automation risk. We took those methodologies, we built an AI-based classifier that used those methodologies to basically take every job in our marketplace and decompose it into its subtasks, and then graded the tasks on being automatable or not with AI. As a result of that analysis, we were able to measure that 10% of the GSV on our platform does have subtasks that are between 80%-100% automatable with AI. That was a new stat we were able to provide that gives us an idea of the total automation exposure in potentially the near to medium term. Although, to put it in context, we only lost about one point of that in the last year.
A year ago, it was about 11% when we did a back-tested version of that evaluation. Now it's down to 10%. Again, that volume keeps declining in terms of its absolute size. We did see in Q1 accelerated automation, I think, on the back of some of the improvements with Anthropic's tools and others. Customers have told us, consistent to your question about what are they doing on the low end, they are definitely leaning harder into automation for the small jobs. Think of it as $500 and below is a good way to think about it, and that's consistent, although this measure gives us a more complete picture of that 10% exposure. Which, by the way, does represent 15% of contract volume because, again, these are the smallest jobs in our ecosystem, typically, that can be automated.
At the same time, we're seeing a lot of AI tailwinds. It's really just a matter of time for those to overtake the substitution that we're seeing at the low end of the market. We have an AI business, so to speak, of category, six different categories of AI work that currently is trending at above $300 million annualized run rate, and it's growing 40% year-over-year, which is even an acceleration from its growth rate a year ago. That's a fantastic indication of how these businesses are really coming in and finding Upwork to be the diffusion layer for how they can get access to AI talent and evolve their businesses. This is, I think, early innings on just a massive opportunity for us.
The natural question is, if it's seeping into the bottom 10%, why isn't this going to keep going up the ladder?
It certainly can keep going up the ladder, but the good news for us is there's a high refresh rate of work on our platform. As long as there is net new work that humans need to be doing, that work does come to Upwork, and it shows up pretty quickly. As time goes on, even as previous legacy or historical work gets automatable, there's all kinds of new things, which I think that the best proxy we have for that right now is these specifically AI categories of work, where the work is being demanded from humans on our platform because of this technology. That's really an undercount of the volume of work that's happening around the edges of what AI is generating. Again, it's early.
You can imagine that automation will happen to legacy job types, there's growth happening with all these new job types that show up and are already showing up as AI unlocks new roles and new experiences for businesses. This is very consistent with past trends. We know in the labor market broadly, I think 60% of jobs that exist today did not exist in 1940. On our platform, it's basically a microcosm of this, where there's just this rapid refresh rate and new stuff is happening. Again, in any given quarter, the deceleration and acceleration can show some volatility. When you step back and look at multi-year views, our platform, the good news is we're not contingent on any one category of work or any one skill. Our platform flexes to wherever client demand goes and the talent responds.
We're really well-positioned to just continue to serve wherever the labor market is going, wherever work is needed.
Could there be an initial shock where people start to experience like I'm experiencing, oh, I can build a card for my wife for Mother's Day in ChatGPT, but then I'm like, "You know what? This really isn't that good. I need someone to come in and put the family in, do the editing, do all this." Is there an initial shock and then all of a sudden everyone says, "I need to take it to the next level," so maybe they're going to-
Yeah
circle back?
We hear that very clearly from customers. The guy this morning was one example. He said, "I could do a bunch of the marketing in theory on my own, but I don't have time, and I'm not a marketer, and so I need someone who's using some of these tools to actually do that work for me." We hear loudly from customers that they want AI tools helping deliver their work, but they want a human expert over the top of that, both because of time and capacity, but also because they want someone who can really quality control the output and make sure it's the thing they really needed.
To that need, we're actually launching a product later this year, which will be our human supervised agent product, which gives our customers access to agentically delivered work, but with the benefit of a human who is overseeing that work, is quality controlling it, is ensuring it's really what it needs to be. That is something we've heard from customers they're very excited about. Hence, we have a product that's in beta on that right now, and again, we'll be launching more broadly in the next two quarters.
You've been here, I think, 15+ years?
Yeah, 15 years.
Everyone asks, what's the magic that surrounds you that has kept you so excited about the future of what's going on?
What's kept me at Upwork, there's probably two things. One is this business has such a huge market opportunity. To see a $650 billion enterprise market on the one hand, a $530 billion SMB market, both of which we are just scratching the surface with. Frankly, in the last year to 18 months, have new offerings specifically tailored for those customers, and also built off of all of the learnings of 15+ years of seeing what these customers need, what works and what doesn't work for them. That is incredibly compelling to me. Right now, there's another layer to that, which is we are playing a pivotal role for businesses as they navigate this AI transformation. If you look at OpenAI and Anthropic and Google, these companies are investing in these deployed co-models that are very focused on enterprise.
They're really trying to go after these enterprise budgets only in terms of helping them with forward-deployed engineers and the solutions they need to really navigate the transformation to AI. No one is doing this for SMB. Our platform is already there as a known high-quality SMB-related solution, which has helped these businesses go through the transformation to the cloud, the advent of social media, all mobile device stuff. When I joined Upwork, everyone was building their mobile app for the first time.
We've always been the destination for SMBs that are trying to navigate these critical tectonic changes and stay relevant. I'm just so excited that there's so much more we can do. We're at the early innings of doing that with some of these new products, with AI agents coming to the marketplace, with all of the things we're doing around both AI enterprise and SMB.
I know it's painful to go through a change in the team, a 24% reduction of force is not a small number. Can you just talk about what's happened?
Yeah. We've been really on this journey around profitability for several years now. If I go back to 2020, which feels like more than six years ago, we were investing a lot in growth. We had a lot invested in our marketing spend. We were doing a brand marketing investment to really see if we could raise awareness and drive that through our funnel. Starting around 2022, it became very clear the environment was changing. Economic buyers out there were being a lot more conservative with their labor budgets. The labor market post-COVID was very different than previously. We have been on this journey of really delivering on the profitability goals that we knew this business could have. It's a highly profitable business, 80% gross margins, a fantastic digital acquisition model, et cetera.
That has involved both reducing our headcount over time as we become much more efficient at certain areas of both innovation, marketing spend, and deployment. Our sales team, we've really retooled also over the last few years. We've done a number of things to really right-size our cost base on the employee side. Of course, today, that includes a variety of things that's partly is AI automation, but certainly that's a minority of the story for us. Other things that really help us on our cost base are the fact we use our own product very heavily. About 3/4 of our team is Upwork freelancers. That's an advantage we continually lean into. Of course, we've also invested in longer-term efforts around our processes, our workflows, reducing management spans and layers.
I would say in 2024, which is the last time we did a headcount reduction, we saw on the heels of that our teams were more focused. They were moving faster. They were more agile. It really actually helped us from a delivery standpoint to have a smaller, more focused team. With the reduction we just announced, yes, super challenging to share that message with employees and say goodbye to some really talented people. We were really building on the lessons from our last workforce reduction, where we saw that actually smaller can be better, and smaller teams can do so much and are not slowed down by additional noise in the organization. That was some of what went into this latest change. I'll say, as we did that, we really were careful to protect the key investment areas of the business.
We did not touch $1 of our marketing spend in terms of performance marketing and everything there, because that's really working for us. We are investing heavily in areas around our enterprise business. As I mentioned, we didn't take any dollars out of that business with this reduction, as well as protecting our SMB and AI-related investments, which are so critical for the next frontiers of the business. There were places where we could trim, and so we did that to really deliver on our promises around profitability and ensure that we could do that even in a year that may have a more volatile kind of environment and top line than we'd expected.
Okay. I guess the last time you went through the reduction, you had 850 heads, then you went to 630. Now you're 480. Everyone's like, wait, just a few years ago, you had almost double the headcount. Is this because of AI advancements? Is this because the market's shrinking? Obviously, it's not the market shrinking, but how are you doing this? Is it AI internally that you're using?
I mean, certainly AI is part of it, but I would be AI washing the story to say, "Oh, this is AI." Again, there's places where we've built finance-related automation. We have an internal product called Talk to Data, which lets anyone query our database, and of course, that means we don't need as big of an analytics team. There are areas where we've had those types of investments really pay off from an AI standpoint. Our teams on the product engineering side, they're definitely smaller because we're doing more with Claude Code and other tools. I would say this has been more a function of like constant discipline and grinding out places where we just see redundancy or a workflow that spans 10 people, and we realize that's super inefficient, we can collapse that.
Having managers who either don't code on the technical side or are not doing a player-coach model in other functions. Those are places where we've really leaned into an operating model that is very lean and very efficient and just doesn't have extraneous priorities or activities around it versus singularly focused.
That showed up in the margin, -1% in fiscal 2022 to 29% in 2025 and guiding to 33%, so.
That's right. We had this 35% long-term target that we're going to hit this year, two years early because of the level of rigor that we've put in here.
You think that somewhere in the low to mid-30s is sustainable over time?
Absolutely. Yeah. I mean, this business can definitely do 35%, then I think there's more over time that we're able to do. Because it's also not just the cost management front. There's cost management side, then there's the growth of our higher margin products, which is where we're seeing our fastest growth rates. Enterprise, SMB, those are both higher margin products. In the last 4.5 , five years, we also built our ads and monetization business basically from zero to now it's more than $100 million business, and that is obviously very high margin revenue. In a tough environment like this one, on the one hand, a lot of the very small clients are spending less. On the other hand, talent is actually willing to spend more for these work opportunities, and that is a part of our platform where there is monetization.
There's a lot more we will continue to do on all of these fronts, higher margin products and cost optimization.
What are the big tools you're using internally that are really helping you on AI? Is it?
On the coding side, I'd say it's Claude Code. That's definitely a big one. I think frankly, Claude Cowork, since that came out, has been pretty mind-blowing to a lot of teams in terms of what they can do now, in a matter of hours and without the level of resource intensity that we would've had before. Those are probably the biggest two. I would say our whole product and engineering stack now is this integrated AI layer that uses Linear and Claude and a bunch of other bespoke tools that give us a lot of efficiency there. Then I'd say in other areas, like customer support, we've had a great deal of success there, as have others.
I think 70-plus% of our customer support inbounds are now resolved by an agent. We know we can bring that number higher because the tools keep getting better. We keep integrating more of our internal data and customer-specific data into those flows so that those agents can be more effective. Those are some of the ones I'd say.
Is that from a service customer, like a service software company, or have you built your own agents?
No, we use a third party called Forethought, which was recently acquired by Zendesk. We've looked at should we build something like that ourselves, because certainly we have the technical chops to do it. The debate is always, is it worthwhile for our precious engineering resources to do that versus use a third-party tool that may be pretty cost-effective? Right now we're using a third party.
Okay. The enterprise strength, what's still the code to crack there? We're leaning on a lot of third-party vendors. I thought it'd be hard to get it through. We're using a lot of expert networks.
Yeah.
We're doing a lot. I initially thought it'd be harder to get it pushed through, but I think everyone's got a pretty open mind about it now. Where are the pockets of enterprise, like here are the next engines of growth that you think?-
Yeah
are coming?
Probably the biggest engine of growth for us on enterprise is with this retooled product that we just launched in the last few months, really, and are ramping in the back half of this year with our first customer migrations in June onto this product, Lifted. What this does for us is it really opens up the addressable market for who we can serve. Previously, when we were going into a large enterprise, and by the way, we're targeting customers who spend about $100 million to a $1 billion+ on contingent work. That is our ideal customer profile for enterprise, and that's about 2,000 companies in the U.S. If you look at those companies, about 5%-10% of their spend goes towards independent freelancers or contractors, and the balance, 90%, 95%, goes to other contract types that are typically deployed through staffing agencies.
It's temp staff augmentation models, and employee of record models, et cetera. Now that we can serve that other part of the market with our new capabilities, that just changes the game completely because not only can we bid for that work, we can also go into enterprise and say, "Look, you're spending $100 million on this program." Because we can optimize both the contract type for the most efficient, lowest cost contract type for the work and the geographic location of that worker to give you the cheapest, high-quality person for that program, we can save you 10%-30% on this $100 million program. That is a very compelling pitch for these businesses, and that's why we're seeing so much pipeline acceleration.
We also were named by Everest Group into their top right star performer quadrant for the solution we offer. It's hardly even in market yet. The feedback we're getting is extremely strong, and that's going to be, I think, a huge growth engine for us because it's just a share winning game. We don't have to go create a new market. This is spend that's already happening. We can do it better, more efficiently, more effectively, and at higher quality for these customers. The other engine for us is broadly what I'd call fractionalization of work due to AI, where we see enterprises that are deploying these AI tools still need humans involved. Those humans do not necessarily take the shape of a full-time employee.
We have a number of solutions that really help us serve them in those needs, and I think that's going to drive a broader trend in the market where full-time employees make up less of the enterprise workforce budget, and contracted and flexible labor becomes a bigger part, and we are perfectly served to win a ton of that business as well.
The stocks had a tough run. What do you think? What are you doing about it from a buyback-
Yeah
From a board perspective, how should we think about this?
Yeah. We are staying the course on our capital allocation strategy. We have an open authorization from our board for a buyback. We did over $108 million worth of buybacks in Q1, which was more than we did in all of 2025. That's going to be an ongoing part of our strategy for sure. We're focused on executing the things we can control, which are building these three strategies around SMB, enterprise, and AI that are showing their performance. They're showing that they're working, and we just need to scale them up to be a bigger part of the total mix, and that mix shift is in flight. We're just staying very focused on delivering the plan and the strategy that's working. We know the stock price will catch up.
Okay. Anything that you see that we can't see from sitting on the outside? What's the most important thing that.
Yeah. I think the most important thing is the AI-related tailwinds that we see. Looking at the categories that are growing on the platform, the type of work that's really changing in the ecosystem and is showing up on Upwork as business demand changes, it's incredibly exciting. Yes, while there is automation happening in certain parts of the business, if you think about a +$300 million run rate growing 40%+ year-over-year in just that part of what we're doing, and that's before we layer in things like our AI agent solution with human-supervised agents. That's before we get the benefit of our ecosystem strategy around our ChatGPT app and other apps that are launching inside these AI-related tools and bringing us new demand channels. There's so much happening there that I think is really exciting and in the early innings.
We have an AI data opportunity that is also very much untapped because we are both the environment for, and the generation engine for a ton of data around work that all of these companies that are building AI agents and foundation models really want. When we look at all of that and the fact that even enterprise, 20% of our pipeline is AI-related projects, we're in the early stages of just a massive opportunity opening up in front of us, and I think it's clear that that's not fully appreciated in the market right now.
You're in the Bay Area.
Yes.
I was with an exec last night. He moved from the Bay to Austin. He goes, "I want to come back. All the talent's in the Bay now." Do you feel like there's this moment where we're seeing? I live there too, so I'm excited.
Yeah.
Everyone's like, "What's it like to live in San Francisco? Isn't it terrible?" I'm like, "It feels like things are better.
Things are better. In some ways it feels very one-dimensional in the sense that you get off the plane in SFO, you get to 101.
10 AI signs. Yeah.
Every billboard is AI, AI, and you're like, "Oh, I know where I am." It's very clear. It's exciting. I think there's a lot of innovation. I think there's a lot of hype and a lot of misplaced fear about the labor market and other things, and that impact from jobs. It's interesting to see Dario and Sam Altman both walking back their comments about AI disruption of the labor market. Now they're both saying, "Look, it's actually not going to be bad." Their tune is very different, and I think that's probably because the data just isn't supporting some of these outlandish claims that they've made. It's an exciting time. There's a lot of innovation. People are motivated. Our team feels it, and we're at the beginning of the next era of work, which is going to be very different.
Thank you, Hayden, for coming. Really appreciate your support of the conference.
Yeah
Good to see you.
Good to see you. Thanks for having me.