Dave Koning: Thanks for all of that. Maybe just one quick follow-up, the number of clients, I think was 502 ending the quarter, I think that was one of the strongest quarters of gains in clients we’ve ever seen. Is that a good early indication that some things are going in the right direction, or is it just much smaller contracts per client on some of the new ones, and that’s why we’re not seeing it yet?
Guo Xiao: I think it’s both. First of all, we have definitely been investing heavily into additional sales and marketing capacity. We definitely believe that we have the right [indiscernible], the thought leadership, and our brilliant technologists, but I think that we have opportunity to expand our demand generation capacity, especially in outbound demand generation. As Erin mentioned earlier, over 50% of our new wins in Q4 came from outbound efforts compared with historically it was only 15%. Combined with our inbound, the outbound, we’re definitely reaching out to more clients or potential clients and we’re seeing acquisition of these new logos and then new clients, and we’re retaining many of them. But at the same time, as you mentioned, deal size tends to get smaller in the current macro environment, and initially when clients start working with us, they don’t sign up as a couple years ago, much bigger deals, so we tend to see smaller deals, shorter deals, but we’re really happy that we’re expanding our client base that gives us opportunities to work and expand from there.
Dave Koning: Got you, thank you.
Operator: Thank you. Our next question comes from Matthew Roswell with RBC. Your line is open.
Matthew Roswell: Yes, good morning. Two questions – I’ll get the easy one out first. Erin, how should we think about free cash flow in ’24? Are any of the restructuring charges, are they cash related, so should we expect like a severance coming out in first or second quarter?
Erin Cummins: Yes, so the restructuring charges did impact 2023. There will be some impact in 2024, but it will be less than 2023. If we compare free cash flow year-over-year, the restructuring charges overall will be a benefit. From a CapEx perspective in 2023, we did have lower than historical levels. As a percentage of revenue, CapEx was under 1%. For 2024, we’re not expecting significant changes in CapEx, but I do expect an increase from where we were in 2023. On the whole if we look at cash flow 2023 to 2024, not a significant change, but given the focus on our operational efficiency alongside the reduced restructuring charges, I am expecting increased free cash flow in ’24 compared to 2023.
Matthew Roswell: Thank you. A bigger picture question, I guess more for Xiao, when you talk to clients, what percentage of them do you think can actually implement gen-AI programs now versus having to put in place the prerequisites, so do they have the data, do they have the architecture to actually implement gen-AI?
Guo Xiao: Sure. In terms of percentage of them who are really already implementing gen-AI at large scale with all the prerequisite platforms, the data ready, it’s very small. What we’re seeing is a majority–first of all, a majority of the clients we’re working with, they’re doing proof of concepts, just trying to prove out both the business case and the technical viability, so I think 2023 was just a year full of POCs. Most of them at the same time recognize that they have to do a lot of work to prepare themselves – we call it gen-AI readiness, and that is having the right data in the right place, with clean data, with updated data, and then building the data platforms they require to–especially the modern platforms that allow them to run gen-AI on top of that.
Only a fraction, I would say probably less than 15% of the clients have these data platforms ready, which was a result of years of working, especially in the last few years, of preparing and getting ready for–not necessarily for gen-AI, but just generally getting ready to get value from data. I think they are running–this small percentage of our clients are running real large scale gen-AI applications using large language models with fine tuning, for example. We do believe that in 2024, we will see more clients moving from POC to real large scale implementation where they’re ready, but most of them will still be doing work to prepare themselves for AI readiness.
Matthew Roswell: Is there a disconnect between that preparatory work for gen-AI and the focus on near term return on investment that your clients are seeing, because I would think a lot of that work doesn’t have an immediate payback.
Guo Xiao: I think it depends on–it’s definitely a good call. It depends on the type of gen-AI use cases. Some cases would require, rightly so, year of investment, at least a year or two, especially building large language models. It would take a while for that to get the payback, and some of our clients are ready to do that and they’re willing to make that commitment, so those are the ones that are moving forward with that type of gen-AI implementation. But a lot of the use cases are, I would say, low-hanging fruit – they are not involved in the fine tuning or large language model building, they are more involved just using APIs of the open platforms out there or building some custom user interface on top of that. I think the most common use case we’ve seen is ultra personalized experience, which is definitely going to return–provide returns in the near term, but it’s also a limited use case of how gen-AI can be used.
Matthew Roswell: Okay, thank you very much.
Guo Xiao: Thank you.
Operator: Thank you. Our next question comes from Kate Kronstein with William Blair. Your line is open. Kate, if your telephone is muted, please un-mute. Our next question comes from Paul Obrecht with Wolfe Research. Your line is open.
