Ben Fletcher (00:11):

I'm Ben Fletcher and I'm really excited to be here today with Shensi Ding, the CEO and Co-founder of Merge. So thank you for being with us today, Shensi.

‍

Shensi Ding (00:20):

Thank you for having me. And also thanks for investing in us.

‍

Ben Fletcher (00:23):

I know you had many suitors and you had a lot of people that wanted to invest, so also we appreciate you choosing us. Well, today in our series we're going to be talking about AI and we are really fortunate that we get to work with a lot of the leading AI companies, but also a lot of the companies that are powering and making it possible for this enablement and this new layer of AI. But maybe we kind of start back and go to the beginning and you can tell us about yourself. Tell us about the founding Merge in the early days.‍

‍

Early days, and the founding of Merge

Shensi Ding (00:48):

Well, I actually met Gil, my Co-Founder freshman year of college. We both studied computer science actually here in New York, and we were in all the same classes, group projects, social group, and Gil was pretty much a prodigy software engineer. I believe he actually got a cease and desist when he was 16 from Facebook. He was that guy. But yeah, he was just a really good software engineer and I was just, okay. And also when we first got started, we were just two kids and everyone was like, who the fuck are you guys? Right?

‍

Ben Fletcher (01:16):

Well, getting started, I always think it's a funny story, but you were the chief of staff at Expanse. You working very closely with the CEO, you ran into this problem, you knew Gil, you all were talking, and then you one day just quit your job, didn't tell Gil about it and just quit and said, we're doing this. How did that go down and what was Gil's reaction?

‍

Shensi Ding (01:34):

He was kind of like.. the fuck? I was like, I'm serious about this. I really think this is a good idea. We had already been doing, I think it was six or eight months of research. We knew that this was a problem. We had the outline of what could be a potential solution. It definitely wasn't perfect or all the way there yet, but I was like, I really think something is here and it's going to keep getting worse and we can really solve this because we have these experiences that are not exactly overlapping and are quite complimentary and we're also best friends. This is going to be really fun. We're going to have such a good time. And he was like, I really wish you had talked to me about this. But then he quit and he definitely doesn't regret it now.

‍

Division of labor and what it’s like working with your close friend

Ben Fletcher (02:15):

I would say, what are the pros and cons of working with your best friend? And you all are very close.

‍

Shensi Ding (02:19):

Yeah, very close.

‍

Ben Fletcher (02:20):

And I would say over time, you've gotten to the point now where you'll focus more on sales and the commercial and Gil's been doing more on product and engineering, but in the early days you were writing code, Gil was out selling, he still sells, you're still writing code. So how has that been working with a best friend and then divvying up the responsibilities?

‍

Shensi Ding (02:39):

We were very lucky that in college we were actually engineering student council, class president, and vice president. So you've actually worked together before.

‍

Ben Fletcher (02:46):

Why didn't you start the fundraising pitch with that? It may have gone a little bit better. I didn't know that you were royalty. 

‍

Shensi Ding (02:53):

I know. Yeah, it was very core experiences in our lives. It definitely should have been top of a resume. Obviously you can't really compare planning parties to planning a company, starting a company, but I had that experience and I just really trusted him. Some things that I also knew he was really uniquely good at was that he was a great leader because when we were in San Francisco, I get to know his team, meet the other engineers that he was working with and they all just loved him. Honestly, he probably took more risk on me than I took on him because he's just such a great partner to have and working with your best friend is so fun.

Shensi Ding (03:31):

Some of the things that we've experienced, it's so unique and no one else can understand. It's almost like Gil and I had a child and you can talk endlessly about this child and no one else gives a fuck, but Gil will always care. Even if I'm just like, oh, I dunno, Fred is sad about something. What should we do? Gil cares. Even if I'm like, oh, this button, the color looks a little off Gil cares. I don't know. It's really fun to have someone that I enjoy talking to that I feel like is really funny and understands me and not only cares about my utility to the company, but also me as a person. And it's just made us really, really close. It's such a unique experience to have with your best friend.

‍

Crafting a company culture that focuses on energy

Ben Fletcher (04:10):

What are the things when you set up the company that you wanted to make sure that you had in the culture that you were building and that we're going to continue and be sustained over the years of the company?

‍

Shensi Ding (04:19):

One thing that I think we knew early on was energy. Whenever we interviewed someone, if someone was energizing, we would feel more excited to work and talk to them. But it was also really important to see how does someone talk about previous experiences? Were they a nice person? And also I've had this experience at tech companies and in finance of being a young woman entering the workplace and not feeling comfortable saying things because I thought people would think I was dumb or I really didn't want to create that kind of environment. And so it was very important to me to make sure everyone was very supportive, really nice, but also direct. We're still very direct, but just really trying to make sure that we do it in a way that shows that we care about that person too. That's why I love hearing this from new people who join the company, that they really feel the culture, that they feel really comfortable asking questions. They feel comfortable highlighting things that are bad processes. They feel comfortable highlighting bugs because it's really important for people to be able to say those things. Others we're just not going to be able to progress.

‍

Why Merge exists and how AI companies are using it

Ben Fletcher (05:19):

Let's talk more about AI. I think it's one of the biggest platform shifts that we've seen and one of the biggest trends that we've seen in our industry. I'm always so impressed with how many AI companies are built on Merge. Maybe talk about how you all built Merge and then how that's helped to power the ecosystem.

‍

Shensi Ding (05:39):

We built Merge to really be the infrastructure for all B2B companies. So every single time a B2B company is built or already exists and you need integrations, you should just be using Merge. There's no reason why every company should be writing the same code base over and over and over again. And that's no different with AI companies. A lot of these AI companies are for B2B use cases, and I'm not going to lie, when we first saw all these companies signing up, we were like, what's going on? Why are all these companies getting built? A lot of them didn't have full landing pages yet either. But I think over a few months we started really seeing a trend of a lot of really great ideas signing up for the platform, using ChatGPT and building things that we would also be interested in using ourselves. And we realized that there might be a really interesting journey for Merge to not only power the bread and butter of B2B companies, but B2B SaaS companies, but also this new wave of AI powered B2B SaaS companies too.

‍

Ben Fletcher (06:36):

And how are those companies using Merge today? Walk us through some of the examples if you can, some of the logos and what they're doing. I always think it's fascinating. You all started building integrations for HR, but then from the beginning it was always we're going to go into all these other verticals. And so you all do ERP, you do ticketing, you do a number of others. So if I'm an AI company and I'm thinking, okay, I need to pull all these other things into my product, that was the limitations in the early days of ChatGPT, it couldn't interact with any of these other apps. And as we're now moving into enterprise, all of these companies need to integrate with your existing stack companies that are out there, maybe some of the new companies that are being built. How have you all built that, powered it, and what are some of these examples?

‍

Shensi Ding (07:18):

Guru is a customer of ours and Rick Nucci, my hero, also the CTO founder of Boomi, he built a great company and what they do is they provide an internal knowledge base for workers. And we are really excited to find out like, oh, we launched this new category file storage, so integrations like Google Drive, Dropbox, Box and their team raised their hand. They were really interested in testing out our new category. And the use case was to be able to quickly pull information from Docs and knowledge-based resources so that if any workers or employees had any questions, they would be able to ask Guru or search through Guru and be able to quickly find the answers. And of course like Gong like very popular sales ops tool. They've always had AI in their products. 

‍

Serving AI incumbents and upstarts, and the verticals they come from

Ben Fletcher (08:06):

Is it mostly companies that were existing companies had customers and they're saying, we see this movement, we want to move into it, we want to launch our generative AI product? And they were adopting Merge or were you also getting the upstart first company? We're building with Open AI and GPT3 or four and we're building our first iteration of the product and we're going to Merge. Which has it been?

‍

Shensi Ding (08:28):

It's honestly both. Yeah, it's really both. I'm just listing logos that I feel like people will probably know. But yeah, I mean we've seen so many every day we see dozens of seed stage companies .AI, they'll use every single category because every single piece of data that you can import from your customers vendors can potentially be helpful in providing insights. So that's been very interesting for us.

‍

Ben Fletcher (08:49):

Are there any specific verticals? Because when we've seen it, there've been the horizontal companies that are going after I would say any task that's out there. And then now we've been seeing more verticalized. So whether that's the legal vertical or whether that's the financial vertical, we're seeing a lot of these more verticalized models that are now popping up in the AI sector. For you all, is it you want to get some of the vertical or is it start with one and then expand to the others or is it every company?

‍

Shensi Ding (09:15):

It's both. I'm not picky. I've definitely see the sales ops companies that mostly integrate with CRM and file storage in order to pull that information.

‍

Ben Fletcher (09:25):

Lot of sales ops, a lot of file storage.

‍

Ben Fletcher (09:27):

What the AI companies are using. 

‍

Shensi Ding (09:30):

Historically also ticketing for productivity. A lot of times people want to see what tickets have been completed for engineering,

‍

Ben Fletcher (09:34):

So a lot less on HR or applicant tracking, whereas that's where you saw some of your earlier customers.

‍

Shensi Ding (09:40):

We've seen that too. Yeah, we've seen everything.

‍

Merge’s top use-cases, the staying power of AI companies and long-term predictions for the field

Ben Fletcher (09:44):

Okay, gotcha, gotcha. What are the one or two top use cases that you like to see when you know that they come through the funnel? It's an AI company. It's like, yes, we are perfect for that.

‍

Shensi Ding (09:54):

Honestly, every use case that's like we need to pull every category because I think in those situations those companies would not have been able to exist without Merge. That's such an amazing feeling because previously if you wanted to build that kind of company and Merge didn't exist, there's just no way you would've been able to build hundreds of integrations in-house without Merge. Now Merge, and combined with ChatGPT and all these other competitors, the world is your oyster. You can really build anything. So I think that's my favorite because there really is nothing else out there but Merge.

‍

Ben Fletcher (10:25):

So you're getting pulled into a lot of these customers, a lot of these startups are building on Merge. Is your view that they're all going to be coming back down? Is this a fad? Is this a trend as you mentioned, every company that's starting as .AI or is it chatGPT 4, XYZ. What's your view on where it goes? Are these going to be durable companies? Are they going to be longstanding?

‍

Shensi Ding (10:47):

Whenever a new great idea emerges, there's going to be 10 different competitors. It's the same thing now, but I think some of these ideas are just really objectively great. I've noticed in sales ops, there are so many different ideas that were not possible previously without AI. And we've also been very interested in using those tools as well. And sure there might be three companies that are doing the same exact thing, but there's going to be a winner and there really is true value that's getting added. And I think it's probably going to be a little bit less like, oh, this company is an AI company, so they're going to get like a hundred X multiple. I think it's going to be more like AI is layered through everyone, but overall I think net for the entire industry, everyone should be using AI.

‍

Merge’s homegrown AI products, and the capabilities they unlock

Ben Fletcher (11:26):

So I mean one of the things that we've always loved is the picks and shovels businesses that are really able to be durable and withstand market trends and market timing and continue to persist over long periods of time. And it's been incredible to see Merge be that for the AI industry, empowering a lot of it. But you all also have launched your own AI products. Maybe talk about those and what the shift was to powering the industry to now having the integrations set up by AI and just walk us through how you're now launching AI products.

‍

Shensi Ding (11:57):

We didn't want to do it just to do it, and I think we had seen a lot of companies that had just launched something and then it clearly didn't work or clearly didn't do anything. And we wanted to make sure that if we were going to be investing these resources on an already very packed roadmap, that it was really going to be worth it. And Gil had this really great idea where we've always had requests for customers to be able to contribute to Merge’s unified API, and we just weren't there yet. But if we could have AI help contribute to some of the mappings, it would speed up a lot of our development. And It could also be a great way to get to serve the Merge community where our customers could feel like they really helped add to Merge as platform. And so we spent a month on Blueprint, which as mentioned allows our customers or anyone to contribute to integrations and the initial mappings to the Merge platform. And it was awesome. It was really, really cool. And I think a really great part of it was that it showed us what was possible and there's a lot more that we can continue doing there internally. 

‍

Ben Fletcher (12:53):

What were the main things that they could now do with Blueprint?

‍

Shensi Ding (12:55):

Honestly, I think the main thing that it really helped us with not only externally but also internally was showing our customers that we were moving really quickly and that we were continuing to innovate. And I think any company that is not always looking at new technologies, trying to move really quickly on the roadmap is willing to adapt their roadmap to test things out, they're not going to win. It gave our customers a lot of confidence that we were going to keep innovating, we were going to keep moving really quickly and that we were the one to bet on.

‍

Ben Fletcher (13:21):

How much will AI continue to persist in the product and for your end customers that you all will power?

‍

Shensi Ding (13:27):

It's going to be very important. We've already embedded AI-powered search and our documentation and it even generates code snippets for our customers, sick! It's going to be really, really awesome. And we have a lot of other ideas too for issues detection showing where something has gone wrong in an integration, but a lot of it's also going to be internal and we're starting to really train our team how to use it a lot more as well. We can't share all of it totally publicly, but there's a lot of good stuff coming.

‍

Standing out in a noisy market of AI products

Ben Fletcher (13:53):

That's really cool. How have you all been able to cut through the noise? One for the products that you're finding in the AI, but as well for how you all can stand out for the products that you're launching, as you say, and we see it on TechCrunch every week. There's a new company that's launching their AI product. How do you all stand out? How do you cut through the noise with everything that's going on?

‍

Shensi Ding (14:17):

I think it's really important to have a product that's very difficult to build. And so that was the first thing that was very important. 

‍

Ben Fletcher (14:24):

Would you say that would be the foundation for how you built Merge or what you're doing with the Blueprint and some of your AI native products? 

‍

Shensi Ding (14:32):

I think it's the combination of both. I think it's great to think of new ideas that are very cool and that you can fundraise very quickly for or the customers find a lot of value, but if you could do it very quickly, I think everyone else can. I think it's the combination of both. The fact that Merge is a huge pain in the ass to build in addition to all the innovation that we're adding on top, that I think shows a lot of value because it's really hard to build. It's really hard to build Merge, and I think it's a testament to when we are successful, it was not easy.

‍

Assessing customer longevity, and how service changes as you scale

Ben Fletcher (15:04):

When you're thinking about companies that you all want to work with, companies that you want to have on the platform, how do you make sure that you can dedicate time to the ones that you think you're going to be around for a long time? You've signed some very large contracts, some will come in self-serve, you'll start by selling to developers and then over time you'll help grow them. How do you distinguish between the AI companies that you think will be around versus those that may not?

‍

Shensi Ding (15:28):

It's hard. I think before when we first got started, we would do everything for every single customer and that's why we did so well in the early days. But obviously it doesn't scale as you reach tens of thousands of customers. So I think, yeah, we don't have the best answers for this still and we're still figuring it out sometimes. So it depends on how many customers they onboard. If you have a lot of customers, you're probably doing okay if you have a really strong Eng team and if you're giving us really great feedback, that also is a helpful indicator to us too in a less direct way. But it shows us, hey, you guys might also be around because of the strength of your team. And then of course also contract size if you're really willing to pay because you're betting a lot that you are going to grow, you probably have some more leading information than we do. So it's those three factors that help us determine how much we can invest in each customer.

‍

Ben Fletcher (16:17):

And the companies that are not great in AI, is it because they just haven't built enough and they're just all reliant on other platforms, or is it just not the right use cases that they're building towards?

‍

Shensi Ding (16:28):

So I've actually heard less about AI companies not being great if they're built on top of other infrastructure. I've just heard some infrastructure products are not great because it's hard. It’s hard. And what's even harder is that now you have such a big audience because there's so many companies getting built right now, you don't have enough time to really battle test your product. So it's like you just emerged as little baby and then now you're forced to compete in the Olympics and it's tough. So I think that's the issue. And I am always very grateful that at least with Merge we had some time to get to that point versus now a lot of these infrastructure companies, you're just getting hammered and you might not have the people or the infrastructure to handle it.

‍

How Merge responded to peaking customer demand and interest in AI

Ben Fletcher (17:06):

How do you make sure that you all are prepared for that? These waves that have been happening, and I would say some huge spikes earlier in the year where companies were coming on, how do you make sure that you're prepared and you can support those?

‍

Shensi Ding (17:17):

So we hire the best engineers in the world as Gil always likes to remind our team, and it's just really important to hire people who care, have the experience, and that we are always planning ahead. So Gil has always been really good at thinking as an engineering leader, what do we need to build for? If something's off, why is it off? And how do we permanently fix this versus just having little band-aids. And so it's really Gil and Gil thinking through this, thinking through what he would want as an engineer and seeing, he sees around the corners, which is really great.

‍

‍The explosive growth of AI as a category, and drawing product inspiration from your users

Ben Fletcher (17:51):

It's been pretty crazy to see the market and what's been happening with AI. It happened extremely quickly. Was it something that you all were expecting?

‍

Shensi Ding (17:59):

I will say I wasn't super surprised by the velocity of companies getting built on it because you've probably seen this too, but every single time there's a category that's clearly a good idea, there's going to be 10 companies with the same exact idea that pop up.

‍

Ben Fletcher (18:13):

You all had a pretty interesting insight into it, right? Yeah. You got to see the earlier companies when they were starting for you all. When did that start? Was that end of last year? Beginning of last year, earlier this year.

‍

Shensi Ding (18:23):

Yeah, it was end of last year. Pretty quickly we started seeing some really small companies signing up and then we started noticing copycats of those companies and then we started noticing a lot of copycats of those companies. And I think copycats are not unique to now. Whenever there's a good idea, a lot of people think of the same idea at the same time, and then they all get fundraising. Fundraising is especially at the very early stage, it's a lot easier now.

(18:46): So that wasn't super surprising. It was concerning though, having a lot of companies that may not last signing up for the platform and contributing to ARR, but that was something that we would have to then adjust for as a business. But I think in the end, as a founder, there's always going to be competition. And that's the beauty of this era. There's so many amazing competitors out there that you see and you can see them getting better and better because the other one is there. And so yeah, for us when we started seeing all these AI companies, we were really racking our brains. What can we add to our company? Seeing all these companies get built on Merge, what are we missing because we don't have any ideas right now. And we had a few quarters where we were just like, yeah, that would be cool, but would that really move the needle and is really worth moving our roadmap out for what a tangible thing that our customers we know they want and we know can generate revenue and add value versus a thing that's a bet and may or may not pan out.

(19:44): And you really pushed us. You're like, you guys really need to start thinking a lot more about this. And I think that helps us a lot to really have an outside voice because it was just me and Gil thinking through like, oh, I guess it's worth it. Do we do this? But I will say in the end, I have noticed quite a few companies just not add AI because they didn't feel like it was authentic to them. I think that's also important. You should be pushing yourself because you don't want to be that loser company that never innovates, but you also can't do something just because everyone else is too, if it's not the right thing for your company. 

‍

Balancing AI and human involvement at Merge

Ben Fletcher (20:14):

Did you all ever have the like, oh no moment. AI is going to take over our whole company. It's going to be able to build integrations on your behalf, it's going to be able to power a lot of what we power today. Did you and Gil ever go through any of those moments at the company?

‍

Shensi Ding (20:27):

So when you get later stage, you meet a lot of VCs and a lot of VCs will just text you random shit. And some VCs are just like, oh, what do you think? I heard that ChatGPT is going to just destroy integrations. And then we would test it out and be like, this is not that level where it can really replace Merge.

‍

Ben Fletcher (20:48):

What about long-term?

‍

Shensi Ding (20:54):

Of course, we were thinking long-term, what do you do to avoid that? And I think in the end, especially every day when you're seeing not just how hard it's to build the product initially, but the maintenance and the operations as a company, we were just like, yeah, human involvement is still pretty important here because a lot of these decisions that we're making are not objective. They're very, very subjective. And even the amount of existing inputs that we have sometimes do not help us. We just have to think for the future.

‍

Preparing your company for surging demand and building products that deliver on their hype

Ben Fletcher (21:24):

Was there anything that you decided to make sure that you could pivot into powering for those companies? And from our perspective, the transformer model has been out for almost 10 years now and the ability to generate whether it's code, or text, or images, but we really saw it when OpenAI, they had been quietly building for a number of years and folks were looking for access to GPT 1, GPT 2, and it was really at that kind of GPT 2, GPT3 three where it became widespread and commercially available. And so we saw all of these companies asking us, Hey, could you help us get access to GPT2? And it was really at the beginning of last year, we started to see a bunch of companies and then we had the opening up and the coming out party with chatCPT where it was like everything. And so from our perspective it was, we saw the way we saw it coming, but then just the onslaught of demand that happened post chatGPT was pretty incredible.

(22:18): How did you all make sure that you were there for it? Because I remember at the end of last year when we were asking you all about it, you were like, yeah, all of these AI companies are coming to us. How did you make sure with this buildup that Merge was ready to go?

‍

Shensi Ding (22:31):

We had to be very explicit, and I think that's important for you to do in marketing is like, if you are building for this, you better use us. And people can't read your mind and sometimes they need you to bridge that gap. And so we did a lot more marketing that was like, Hey, if you're an AI-powered B2B company, you need integrations, so you should just use us, make it a lot easier. The great news was, at least at that point, we had had a reputation and we had customer existing customers, and there's a lot of word of mouth, but even still, that's not enough. And if you're even earlier stage at that point, you have to be really, really, really loud.

‍

Ben Fletcher (23:04):

And how much of it is marketing versus really actually powering, right? You see a lot of blog posts and you see a lot of folks that are out there saying, we do AI or we power AI. But how much of that is the distinct decision you make at the company to have the right products, have the right support and be the right product for those companies?

‍

Shensi Ding (23:22):

It's both. So I remember hearing from this founder of a unicorn, he told me a long time ago, he was like, one thing that's really important for your company is to create a discourse that your company has the best product. But the thing is, you also actually have to have the best product to back it up because otherwise it doesn't mean anything. Sure, you can be like, oh, this bowl is for AI, but if this bowl doesn't do anything for AI, then it doesn't do anything. It doesn't mean anything. And people are going to quickly see what's real or not. 

‍

Understanding customer wants, and knowing which categories to target

Ben Fletcher (23:49):

How did you pull in existing customers to understand what they wanted to really build for it? And what were you hearing from them early days when you were launching these products?

‍

Shensi Ding (23:57):

So the great news about some things are that your customers are already telling you, it's not really you having to pull information from your customers. And I actually think that's more important. If you're getting yelled at by your customers to build certain things, then you should probably do it. And that's what I love about B2B is that you never have an open question about whether or not something is going to be helpful because your customers are already telling you.

‍

Ben Fletcher (24:17):

And how loud were those voices from your customers telling you they wanted these products?

‍

Shensi Ding (24:21):

Pretty, pretty loud. And so it ended up really impacting what categories we launched. So previously we had HR, accounting, CRM, ticketing, marketing, automation, and we sped up file storage for some customers like Guru, Assembly, and a few others that were like, Hey, we're building this AI product and we really need this category. And by the time we had a few dozen customers requesting this specifically for AI use powered use cases, we were like, okay, we need to fucking move. And so we built it really quickly, worked very closely with them, but they really came to us and told us. And then the next category that we're launching soon too was really from a lot of these customers as well.

‍

Ben Fletcher (24:58):

And then are there other categories that are more amenable or more needed by a lot of these AI companies?

‍

Shensi Ding (25:04):

I think it's combined with what's hot and what's doing well right now. I think with the market downturn, it's hard for some categories to do super well right now if there's not a lot of hiring, it's tough for more recruiting companies to pop up right now. So that category, it is just harder for an AI company to be incentivized to go into there. But for sales right now and customer success and productivity and knowledge work, all those categories are still doing very well. And so there's just a lot more interest and categories that serve those use cases. 

‍

‍How Accel thinks about longevity in the AI space

Shensi Ding (25:40):

So when you were seeing this AI wave and all these companies emerging, did you have any portfolio companies where you're like, they're fucked?

‍

Ben Fletcher (25:47):

We definitely went through the exercise of looking at the portfolio, and I'd say it's a spectrum. I'd say you're looking for more of, I was invested in a company and they do automated sales commissions. And it was like, well, that's probably not one where they're going to be completely taken over by AI. But they did a lot of things that they could incorporate into the product, and it was really, really interesting for them to see, hey, what was the journey and how they get there? And now they have a really interesting product for what they can do. And then I also work with a company that does automated customer support, and it was really, really cool to see how they went from implementing the transformer model and these large language models into their product. And that was over two years ago. They had hired somebody from Genesis and somebody that had this unique background and they could actually build that into their product. So I saw a lot of that coming early on, and they actually got selected by OpenAI to power all of their customer support.

‍

Shensi Ding (26:37):

Wow, that's very cool. So good for them.

‍

Ben Fletcher (26:38):

But it was awesome. But in the very beginning, we definitely had our moment where it was like one of the biggest use cases was going to be around customer support.

(26:47):

And you look at, I would say, where it collides in the areas and just how I would say how much they're exposed to what these large language models can do, and the companies that were really able to incorporate that quickly, they were able to launch that into their products. They've been a lot better off and getting there quicker and being able to capitalize on a lot of the market demand has really helped. Versus sitting back and saying, we're not exactly sure where this is going to go, and we're going to, exactly what you said, we're going to sit on the sidelines and see. They missed a lot of the demand that they could have had if they had leaned into it.

‍

‍

Separating founders who are chasing trends from the true AI-believers

Shensi Ding (27:22):

How do you differentiate founders that are building things for AI because they're really authentically interested in it versus ones who are only doing it so they can quickly start a company?

‍

Ben Fletcher (27:34):

Yeah, I think it's pretty clear when you're talking with those founders, a lot of the best founders that we work with, they have very much felt the pain of that product that they end up building. And so I think about Daniel Yanisse, who's a founder that I get to work with at Checkr, he built the API for background checks because it didn't exist when he was looking for that product that was in the market. I think about Eric Rea from Podium when he was building this product where he wanted to connect the local tire shop with their end customers. He built it for his dad in Canada. And so a lot of the best founders that we get to work with, and I think about Merge, it's the same thing. It's you all struggled with these integrations. A lot of the founders that are building AI native products, a lot of it, it's been their life work, and they're saying, we've been wanting to make these breakthroughs. We've been wanting to launch these products. Now the technology's there, the compute is there, they can actually go out and build these. Whereas you find some of the folks that are going to that because that's where the money is or that's where the market has gone. They're definitely not as passionate. They're not as thoughtful, they're not as opinionated about their end market, and it really shows when they talk about the vision, where they want to go, and really their passion for what they want to build. Thank you for doing this. As we're wrapping up and as we've had the conversation around all these things that are going on in the market, I'm curious for you, what are you excited about for the future of AI?

‍

Shensi Ding (29:01):

AI is not a single player game. There's going to be a lot of different companies that help contribute to the success of this new generation of companies that are getting started right now. And I'm just really excited to see how Merge can help contribute to the ecosystem. I love tech startups, and it is so cool to see all these new ideas that are being generated, especially the ones that wouldn't have been possible previously without integration. So I'm really excited to see what everyone ends up building.

‍

Ben Fletcher (29:24):

Awesome. Really cool. Thank you.