SaaStr GTM Stage

Please take your seats and silence your devices. Please take your seats.

Our show is about to begin.

Please welcome to the stage Global Head of Deal Management, Benchling, Uri Ozil.

Head of CS, Monday.com, Rob Schultz. Jason M. Lemkin. Global VP, GTM Strategy and Operations, Databricks, Sahana Sarma.

And founder and general partner, Recursive Ventures, Itamar Novik.

Okay, hello everybody. Welcome to the most exciting panel of the day. Yes, I know we don't have Snowflake CEO up on stage here, but if you're like me and you care about go-to-market and you're like just going insane over all the tools that are out there, this is the panel for you. So before we get started, who here is a go-to-market professional and is using AI today? Let's do a show of hands. Yeah, look at that, that's more than half the audience. So exciting stuff we're going to be talking about here today, and we'll start by doing a quick round of introductions. Rob, kick us off.

Thanks, Imar, appreciate it. Well, happy to be here. I'm Rob Schmelzer. I head up strategic CS at a company called monday.com. Great. Yeah, I hear some things.

I love it.

Uh, yeah, and I'm really excited just to talk about how AI is impacting our organization and our ability to engage more successfully with our customers.

Thanks.

Hi, I'm, uh, Sahana Sarma. I'm the VP of Strategy, Growth, and Operations at Databricks. Before that, I was at GCP at Google for a number of years, and before that at McKinsey, where I was a leader in their tech practice. Super excited to be here and talk about how we're all experimenting with AI.

Hi everyone, my name is Uri and I work at Benchling, which is an R&D platform for lab scientists, which is SaaS-based. I've been there for about 5 years. Prior to that, worked at Autodesk as they were transitioning into SaaS a million years ago. And yeah, I'm very excited about talking with you today about go-to-market efficiency and the exciting new tools we see around AI and where it could go.

So one thing that's really cool about this panel is we've got 3 people from 3 different companies who are focused on different parts of go-to-market, right? You have Rob is really in the customer success strategic kind of side of things, obviously very interested interesting for enterprise SaaS, right? You have Sahana who's like more generic go-to-market strategy, sales enablement, so on and so forth. And you've got Uri who's on the deal desk side actually getting those deals done and kind of in a supportive role supporting all the account executives getting out there and getting the deals signed. So hopefully what we're going to see out of this mix is a lot of different tools that are being used by different go-to-market teams. Teams and that these guys can weigh in on. And hopefully we'll also have a little bit of time for Q&A for you guys. So let's go. So first question for you guys, and obviously feel free to jump as you see fit. What's, what's your team excited about in AI? Like, where are you seeing significant adoption? And we'll obviously get to the actual tools later on, but where are you seeing more significant adoption across your organization? Is it more with like customer success, with sales, SDRs? I know there's a lot of AI SDR conversations out there, right? Account management, like where are you seeing the most action, I guess, with AI tools within your go-to-market teams?

I can start. I think the first place we're seeing it is actually in the field with our AEs. I think the use case around just reducing some of the toil, right, around creating briefing documents, trying to create account plans. We're seeing a lot of excitement around kind of getting those first versions out there. The other thing we're looking at is areas like enablement. How do we really think about using AI tools to help our field practice, assess, especially when you're giving them all that training? But how do you really give them that opportunity? And we are, of course, trying things like the AI XDR and looking at more pilots in that area. But across the board, looking at different, different types of use cases.

Yeah, I can go. Um, I feel like for us it's mainly around— not mainly, there's also customer success, but mostly around the sales and SDR. So prospecting, uh, essentially pipeline generation supporting tools, um, as well as, uh, data enrichment, which is kind of like understanding, um, all the different facets of information that can funnel into your prospecting of customers. And then I would say on the customer success side, we do use AI mainly for more customized outreach and also for data for customer scoring. So understanding different parameters and understand how they affect the customer, how they affect the relationship, and using that for, uh, for either upsells or just to maintain a good relationship.

Yeah, and I think just to add, add to that, I think the area I think across the board field-facing roles at Monday are seeing early value is taking kind of repetitive stuff off their plate— note-taking, getting summaries put together, generating action items to follow up on using AI. And we use, uh, Gang, Gang, to, to do some of that for us. I think that's interesting. I think what's gotten some of my team excited is some of the path that we're heading down around generating insights. So at Monday, you know, I've really been focusing my CSMs on putting more emphasis on developing more executive-level relationships, which, if they're not typically doing that day-to-day, is a bit of a scary thing to jump into. And one of the ways we're using AI is is to really generate insights that are both at the company level but also at the individual level. So for example, we use Pocus, Pocus, their AI components, both to generate insights about, say, a specific enterprise customer. But we do two other things that build on top of that. One is we've actually loaded in— we've put all of our field people through training around force management. Command of the Message, MedPic, for those of you who are familiar. We've loaded that in as well. So when we generate these assessments of what's going on with their customer, we're also looking for what are the key critical initiatives and outcomes that customer has, and then marrying up what's Monday's relevant value proposition that connects to those outcomes. So when they go into a meeting, they're already pre-wired for that storyline. And they actually have a cool plugin. And I'm not, I'm not selling Focus here, but I'm just telling you what we're doing. They actually have a plugin that goes into our calendar so you can pop those insights in. You have a meeting coming up with a customer and it actually gives you relevant insights for the people who are attending that meeting. So it's actually kind of cool. And that's one of the things that our CSMs have gotten super excited about.

Very cool. So I didn't introduce myself. I'm not going to do that. It's too late for that. But I'm a venture capitalist. And I don't think everybody necessarily knows that venture capitalists have, you know, we also do sales, right? Because we need to pitch our limited partners. Those are investors in our funds, and they need to— we obviously have to convince them to come in and invest in our funds. It's actually very similar to enterprise sales sort of process. So I just say on my end, I'm very heavily relying on prospecting tools now with AI. And also on sort of pre-qualification tools, right? The things that I've been able to show, let's say, a virtual assistant that I used to employ in the Philippines, say, hey, go through this process of like figuring out, okay, is this person working at a family office, right? Are they deploying capital into emerging managers, so on and so forth? I've been now able to automate with really cool tools I just, you know, putting in a random plug here. I just invested in this company called Harmony, Harmony AI. And what they do, it's like, it's like a Loom. It's basically you record whatever workflow you want to have and then it breaks it down to an agent that can repeat that flow. And you can also design and have like all sorts of different considerations when it goes through that workflow that you just basically videoed into. So very interesting times.

Right?

And, uh, you know, moving on to the next question. As a VC, I probably get to see at least 2, if not 5 deals a week— I'm not kidding— that have to do with AI go-to-market tools. It's just crazy, the proliferation of tools that we're seeing across all the different departments. Um, so one thing that I wanted to do, and maybe, maybe we can pull the slide up here so everybody can see, is kind of talk through more specifically about the, the stack that you guys are doing in your, like, using in your different companies and, and sort of talk more about how you're using them. So we got the stack up here, and who wants to jump in and talk about what they're using and why?

Sure, I can go. So like everybody else, we're using Gemini and ChatGPT on a high level And that's mainly for content creation and analysis. I'll take a minute more just to talk a little bit about that because I feel like a lot of the point solutions you see in AI are great, but that's where they are. There are point solutions where with Gemini, if it's integrated to all your other tools and your, your G Suite and whatnot, or ChatGPT, you're able to build more customizable but still got a wide range type tool. Tools. And I'll give an example from my role, which is more go-to-market ops, where we are— we've been spending in the last 2 quarters or so training Gemini to essentially be a deal desk assistant or a copilot, right? You feed it deals, you feed it playbooks, you, uh, while you connect them with your CPQ, and you're essentially saving a lot of time. And, uh, and also kind of like the clicks for different people to get deals out the door that are optimal. And that's something that, you know, you could potentially— I'm sure there are very young-stage companies that are already doing that or starting to do it. But when you can do it yourself using an existing tool that is already adopted and already integrated, it's a great win. So I want to just call that out. We definitely leverage Gong a lot, both for transcriptions and intel on that and coaching. We also use it for pipeline and forecasting, which is great because again, when you synthesize the two, it works really well. And then I'll talk about two other more, the list is long, but the two that are kind of like bringing the most vocal support in our teams are Clay, mainly again because it is not a point solution, but it's integrates with a bunch of point solutions and enriches data, but you can kind of like build it into your own thing. And then Scribe, and again looking from an ops perspective or an operator, the ability to go through a process, have it fully documented similar to what Itamar was talking about, and have that built into a sales enablement course is really impactful and really helpful and saves a a bunch of times. So, um, I'll leave the floor for the other panelists though.

Thanks.

I'm happy to jump in. So, you know, similarly, we have, uh, our field using Gemini, ChatGPT, and Perplexity for sure to do sort of deep research on their customers and understand what's changing. Uh, we've built our own tool, Ask Mo. That's, that's, uh, we've blogged about that, that really provides an entry into Salesforce for us to allow for our AEs to ask a number of questions across our field, like, what's the next best product, uh, to offer my customer? Looking at what are other customers doing and what's similar, uh, looking at, hey, what's— what does the organization of this customer look like and what's changed? Uh, can you create me an executive briefing document where we're able to pull in? So we're using really this agent to pull in a lot of information that we have across our, our stack and as a way to easily serve that up to the field. So that's something we've built. We use Glean quite a bit to build agents on specific topics. Right now we're growing so fast, bringing so many people in. Glean is a pretty effective way for us to get information out to our field that's consistent, especially from a content management perspective. And then Udely and Uplimit we are using to actually drive enablement and allow us to better assess, you know, one of those eternal questions, how does enablement really impact success? So, and, you know, of course we've had things like scoring that you mentioned, you know, using that kind of data to actually score customers, figure out where to engage. So, and, you know, the last piece, we are doing a number of active POCs to see if we can broaden the way we prospect, right? Uh, and really looking at, can we have something called touchless prospecting using some of the tools where we can create those emails to customers that are generated? The one thing I'll say about that is you still need to have some kind of human in the loop. We're not yet, uh, in the point where things can go out customer-facing, uh, without some level of just risk and quality control. But we are working actively on creating that use case.

That's interesting hearing what both you are doing. I, I'll add a couple of other tools. I mentioned Pocus already that we use, that we're using pretty heavily to really run our CS playbooks. We're using— I mean, we're taking advantage of some tools that we already had in our toolkit as their AI capabilities turn online. So Salesforce is an example. We use their agent force to to automatically map our CSM engagements. So we categorize by engagement type and we track all that. And that was something that our CSMs used to have to go in and manually log, and now it just happens automatically. It looks to see if there was a call or an email or some other activity across our systems. Salesforce just captures that, it automatically categorizes and puts them in the system. So really just an example of trying to take manual repetitive things off of our team's plates and using our tools and AI to, to remove that so they can focus on other things. Another tool didn't make it onto the slide, but we have a tool that we're using called Interpret, which looks across our, our customer communications to gather trends around customer product usage. In product needs. So we, we're really good at understanding individual customers and what they need or where their asks are, but it's challenging oftentimes to aggregate that data. And Interpret's been a nice tool that's allowed us to look, look across our customer base and identify trends or themes by segment or by other areas. And it really has helped our conversation that I have and my team has with our product, with the product organization. So another tool we're using AI that's really helped us do a better job of getting better quality of information that we provide to other parts of the organization.

It's funny that you mentioned that. I was one of the— a member of a team that invested in Totango back in 2013, and that was really a lot of the premise back then was we'll help you really track on the, obviously, customer success side how your customers are using all those tools. And guess what? We had to wait another 12 years for AI to really be here, enable, you know, having this kind of wide-arching view of customer activities to really facilitate that on the customer success side.

It's really a powerful opportunity for us as go-to-market leaders to be able to have a tool that can aggregate data for us. It just makes it so much easier for us to go to other parts of the organization and be able to advocate for things or make the business case for things. That was oftentimes something that just took us so many hours of analysis to do. And now we've— we're starting to have tool sets that put that information at our fingertips, which is pretty cool.

Very cool.

Okay, switching gears. So we got all these tools out there, right? This huge proliferation of tools, and all the vendors are chasing, you know, down these guys here on stage. And what I'm hearing from some of my friends who are running go-to-market teams, they're like, Hey, we don't necessarily have more budget. Like we're already paying for all those tools and the AI tools are incremental. Like they want more money and I don't have a budget. So what I wanted to talk about next is kind of the brass tacks of like, how do you guys make those decisions inside your organization on, you know, sort of buy versus build? I mean, you can build some of this stuff in-house with ChatGPT or whatnot with OpenAI. You can, you know, you can choose to buy certain tools, not buy others. Like, how do you prioritize, and how do you figure out if and, and how you're going to spend this incremental, uh, budget to, you know, introduce AI tools into your organization?

Yeah, I think I, I can start, talk a little about our thought process. I, I think we really try to identify where we think are key areas, either areas of need or areas of opportunity in terms of things that are taking our teams a lot of time to do, or things that we're not doing at that level of quality that we want. And I think once we've identified those high-priority items, we look first to see, do we think we have something in-house already that we can do it? I mentioned Salesforce and activating their AI to do, to do a certain task. We'll then, we'll then look into our own internal organization, and we have a internal tool called Big Brain, which increasingly we're using to deliver exactly AI-driven initiatives internally. And then I think at that point, we'll look externally for a point solution. I think it's tricky with point solutions because, you know, one of the things that I'm already seeing in our organization as we've added these additional tools, which we have a number of them, is it makes it hard for our users to know where to go to get their work done. And if we add more tools, they start to say, "Well, wait, I go here for this, I go there for that." And that really is a friction point that makes it hard to drive really good adoption. So I think one of the, that third thing that we think about is how does it fit into how we want our people to get their work done? and we want that tool to be a net positive addition to that workflow, not something that makes the workflow more complex.

Yeah, maybe I'd add to that. I think the first lens we're actually looking at is through the field experience lens, because one of the things we hear is with all these point solutions, things have gotten super complicated to actually execute. So that idea of what actually is creating simplification is maybe the first lens. People don't want to go to 6 different places to use 6 different tools. So that idea of one place, whether it's Slack or whether it's Salesforce, that's a really key usability point. I would echo, where we can, we are trying, trying to build to avoid that kind of proliferation. But where we do see, you know, value, we are certainly, you know, looking to expand. So we're trying to prioritize our use cases as well that way. The one thing I'll say is we do seem to be working in a different operating model on the AI type use cases. Maybe in something more traditional, we'd write, you know, I'd write a requirements document for IT. IT would create options. Because of the speed of what's happening and trying to roll things out quickly, we're working a bit more two-in-a-box, right, with IT. So I think the operating model for these kinds of applications is also a bit different because you can't just wait, you know, for that whole cycle of POCs and bids. But we're trying to be much more thoughtful to the experience because we certainly don't want to overwhelm the field.

Yeah, from our perspective, I feel like the— again, everything I'm— echo everything I'm hearing here. And definitely you try to prioritize your areas of opportunity and your current gaps and go-to-market, or what you conceptualize those gaps to be, and then you try to find the right solution. But I think another layer of it is kind of like shadowing the rep, right, and understanding where they spend their time on what matters versus what's the administrative or the kind of like the support type work. And when you see kind of like the number of clicks and the number of hours that they spend on documentation, prospecting, searching, working CRM, that's where you understand kind of like where you're— how can you get to better deal velocity, better pipeline generation, better lead generation, and just save them their time so they can spend time on the, you know, the $1,000 an hour type work, right? And so understanding where those hard hurdles or challenges are, and then really when you vet those point solutions or, or any type of solution, understanding can you really draw the line to understanding, okay, how is this gonna save them 2 hours or the 1 hour a week or a day they spend doing something that's not gonna help them build relationship or get a customer to a demo. So those are kind of like the lens that I look at those, at least for the field-facing type, type solutions. When we think about more more operational stuff, like, you know, things that would work for a deals desk team or somebody who's working on RFPs or kind of FP&A. Those are all supporting teams that support go-to-market and they do a bunch of work and there's always a pile of other things. I wanna say bad words, but like other stuff they need to get done. And so understanding how can you be more efficient being efficient there is also super critical because it's also a cost-saving exercise.

Very cool. So obviously at this point the vast majority of you, and I know I feel the same, is kind of afraid that AI is going to take our jobs, right? So it's scary times that we live in, and I think one kind of going deeper a little bit on the previous question that I asked is maybe when you're thinking about those AI tools, or were you having internal conversations with your CFO when it comes to budget, etc.? Is the conversation more like geared toward productivity, like we can get much more productivity, or are you starting to think about AI as a, as a potentially a cost-cutting tool, right, within your organization? How's the— how does that conversation go internally?

I mean, I, I think there's certainly— I think right now, given everything, there is certainly an aspect of what efficiency can we drive not just with AI, but any other even traditional tools like automation. But I feel one thing about AI that we are seeing a broader decision on is how much intelligence does it bring to the organization? Can you do something better? Can you make a decision, uh, more effectively because of it? And that is a different kind of value than just the efficiency. And I think we're all seeing that, right? Even on, you know, to the point of operations. Hey, can we make a better decision on this deal? Can we make a better decision on the way we structure compensation? You know, so there's all sorts of decisions where we're seeing AI help us, and I think we're valuing that as much as the pure efficiency side. Uh, so we do see it, but I think it's looked at in a broader lens than just efficiency.

So just, just Jumping in really quickly on that because I love that framework of, you know, potentially increasing the odds of being able to close this prospect and so on and so forth. I think it's very refreshing to hear that. But I guess my question, my follow-up question on that would be, how can you figure out that this thing is working? Like, how do you get to a conclusion? How can you measure and realize, oh wow, we're actually getting better outcomes by using this tool? When it's really hard to compare what would have happened if you haven't had the tool versus what happens when you have it.

I mean, at least for us, we have a pretty robust data science team that's helping us run like the synthetic control, right? If we ran these customers with these kinds of capabilities and this customer without that, you know, this kind of go-to-market without those capabilities, are we seeing any differences? So we do a lot. We've been doing a lot of A/B testing with our tools.

Very cool.

I'll add to that also. I think there's also, again, coming back to the rep, I think there's also anecdotal information that you can, you know, when you do either postmortems on deals won, deals lost, QBRs, you know, in the last couple of quarters you hear the sales rep saying, this is a life changer, right? Or this, I was able to produce Customize emails to all my prospects, all my existing customers. My response ratio is like doubled, right? And they would be the best advocates or the ones who are gonna shoot down tools that are either redundant or really helpful. Doesn't mean that you don't look at data and you don't use the data science folks. It also matters if you're a 10,000-people company or a a 1,000-people company or smaller, but I think you can also get that from the people on the field.

Very cool.

Yeah, I think I just— to add to those things, I think we're probably in the middle there somewhere. If I even just think about our CSMs, I mean, we've set an aspirational goal of what we'd like that role to look like and what we'd like them to be doing, which is different from where they are today. And so part of using AI and using some of these tools is to say, hey, we're going to help you get from where you are today and the work you're doing today to what we believe is higher value work that allows us to deepen our relationship with the customers and at the end of the day, obviously convince them that, that they see enough value that they want to, you know, invest more with, with Monday. So I think we've got a mix of Are we taking the low-value work off our team's plates, and are they now repurposing it to higher-value work? We're tracking those things. And so I think there's that aspect of it as well that we're seeing come into play. And it's not necessarily one tool that's going to get us there. And I think that's the trick. If I turn around and put my vendor hat on, that's oftentimes the trick as a vendor is you might be contributing to one part of that journey, but not the entire part of the journey. And how do you make sure your customer understands the explicit value you delivered in helping them get to what might be a bigger outcome or a bigger goal?

Another way is a point of comparison, right? You talked about adoption, right? And Rob talked about adoption at the user level. And in a perfect world, everybody would be using 100% of the tools you give them, but that's not the case in all places, I would assume. But I will say, when you take a rep that has been— you see the metrics that show that he has been using, he or she has been using the tool or tools, and then you see more of a more stagnant or late adopter rep, and you compare their trends in their performance metrics and their ability to close and ability to to create pipeline and execute on that pipeline, you see, you know, the variance and trends, that gives you an indication of whether that tool is really helpful or not, for sure.

Very cool. Okay, so we have some spare time, so we're going to have fun and we're going to go off script.

Uh-oh.

Let's see how the panel feels about this one. Just a quick round of questions. AI replacing SDRs. You think it's going to happen? Yes or no?

I'm not going first.

A little out of my wheelhouse. So I, I think, I don't know, I think it's, I think, I think the SDR function is going to change over time. And so it's going to be a different world for SDRs in the next few years. And I think AI will contribute to that, but, but completely replace I still have a hard time. I still have a hard time seeing that.

I think not, not yet. I think we're gonna be though managing agents and people pretty soon. I think you'll see a mix of, you know, you're managing 50 agents and 50 people, but I don't think it'll completely replace the function, at least in the next, let's say, 2 years. What timeframe?

90 days.

2 years is eternity in AI.

Yeah, that's what I said. 2 years, I don't see it all going away, but I think after that it's unclear. It's hard to know beyond that given how fast things are progressing.

I think that from my perspective, it's like Rob said, I think hit the nail on the head. It's going to change the function, and if you can't change, you're not going to be in that function anymore. And I think that relates not just to SDRs but any kind of work that's mainly or foundationally there's an administrative portion, part to it, right? And it could be an SDR, it could be a lot of other go-to-market supporting roles where the current or the past way you did your work was mostly just following a checklist of things you need to do. And that will have to go away because again, if you want to justify your existence in that role, your role will have to deepen and you either have to build insights, build relationships, which, you know, AI cannot do today. It can, you know, start a conversation, it can't hold the conversation, right? Or provide some kind of added value that's not there. So I think it's a trick question because if you're looking at SDR the way it was yesterday, yes, that will go away. If you think that, but if you believe that the function can evolve into something more meaningful, then those people who are able to do that and redefine their role will stay.

Very cool. So, so, you know, as a venture capitalist, I try to live in the future. I've now gotten to a point where my SDR, which was also me in the past, is now fully run by AI. So I just reach out to prospects on LinkedIn and the thing has this natural conversation with them, pre-qualifying them. Are you actually an LP that would invest in emerging manager? Yes or no? Great. And then they hand it off to me. So I'm a believer that yes, we are going to have AI replace SDRs. Not everywhere, not all the time. But when you have some sort of a potential deal or transaction that inherently can be done fully digitally, and obviously if you need to meet people in person, that's, that's not going to be an AI, not, not yet until the robots take over. But I think we can do a lot of that especially the SDR side, with AI, and we could do it pretty well, and we could just manage all these agents that do like 90% of the work and we oversee them. And I think the technology is very quickly getting there. I just want to remind this audience that ChatGPT launched what, like 2 and a half years ago, right? So just imagine what's going to happen 2 and a half years from now. One last unscripted question before we move into audience Q&A. If there was one AI tool that could exist, it probably doesn't exist today, maybe it does, but it probably doesn't exist today, that you could get for your team now, what would that AI tool be? And if you don't have a great answer, that's okay, you don't have to answer, but just like, you know, trying to open up our minds to what would be a great future for you and maybe some folks here in the audience can go build that for you.

Anybody have a first thought? I'll, again, looking at it from a deal— from a— my core function which is pricing, packaging, and those kind of things. I'll— In a lot of larger SaaS companies, there's a lot of time and money spent on customer surveying and understanding kind of like the customer's willingness to pay, kind of like essentially trying to build a demand curve for your product, right? How to bundle it the right way, how to package it, what segments to sell it in different ways. And up until very recently, maybe even today, go to high— you would work with a consulting firm, pay them half a million bucks or more, uh, spend a quarter and a bunch of your own analyst time to get those insights, and then, you know, drive towards execution. I feel like we're getting very close to a world where you could leverage either one of the basic— like, not basic, but the fundamental, uh, language models like ChatGPT, Gemini, or, or even, uh, uh, like starting to see some point solutions out there that are helping you use both your existing sales data, customer outreach to essentially pull a lot of really good feedback from customers about their behavior, their purchasing behavior, and then using that to do a lot of, you know, back office math to get you to a place where you have a more insightful go-to-market pricing and packaging. And again, it's something that I look from the lens of my my role, but it— I can see that as a huge time and, and cost saver.

A shameless plug for a Recursive Ventures portfolio company. If you use Perspective AI today, I think you can get 80% of the job done. You can feed Perspective AI with your question of like, what should I price this, blah blah blah, and your data, right? And then you get a link and you can send that to 100 of your existing or customers or prospects and we'll have a natural language interview with them with the questions that you want to have and then return all the answers in a way that you can really kind of slice and dice in a great way to have insights on your pricing. Again, not all the way, but kind of 80% of the way.

It is a shameless plug, but I appreciate it.

I kind of like this. Now I'm going to throw an idea out there and see if you got an answer for me. That's pretty good. Starting down the path that you started down, which is I think on the customer success side or the account management side, we really grapple with getting a good understanding of the overall account map and almost like a heat map of our customer, right? We have better understanding of the people we have direct relationships with and direct product interaction with. And then we've got these big gaping blind spots across the organization. And I would love a tool that could go out and almost pulse an organization and come bring back to me what part of the organization's hot, what part of the organization's not, where do I either need to apply more resources to make sure I'm uncovering an area where maybe there's something going on that we need to pay attention to, and conversely engage with parts of the organization that maybe are, you know, I hate to say ready to buy because I'm in customer success, that's not necessarily how I'm wired, but but receptive to our value proposition. You got that?

Nothing in the portfolio yet. All right. But in the last month, I've seen two customer success/account management tools that focus on helping you prioritize which customers to upsell to and why, based on all the data that you have across engagement, usage, Slack messages back and forth, Gong recordings, everything. And they give, it's actually, both of them were aimed, very similar concepts, Aimee, the head of customer success, would come like, here's your prioritized list of the customers you should pay attention to.

So I want the white space afterwards, and I want the white space in there too. So I want it all, but thanks. That's— we'll talk about it afterwards.

It's, it's definitely happening. Give it a little bit more time.

Um, okay, I think, uh, maybe I'm biased. I used to run the global value team at GCP for some time, and if there was a way way to create something that took very customized data and was able to translate that into a value story coupled with what's the right, you know, pitch, frankly, for a given customer that's very personalized. I think that would be incredible.

I take that tool too.

Yeah, you take that one too, right?

Yeah.

So, yeah, something close. Uh, Perceptive Panda, another portfolio company that's doing great. They do micro interviews with mostly with prospects, but not only. And through those micro interviews, they better understand what are the priorities of this prospect? Like what are they looking for? Who else are they comparing this product with, et cetera, et cetera. So when your account executive actually gets in that first call, they already have all that fielded, right? Like this customer is looking at A, B, and C, and they prioritize this, this feature most, and that's how you can come kind of more prepped to understand what the customer requirements are. Perspective Panda. Okay, it says Q&A out here actually on the slide. Not that you guys see, but it's Q&A time. So we would love to get questions from the audience. What's top of mind from you guys? Just lift your hand and try to—

yeah, get some microphones out there too.

Is there a way to get a mic to you, or—

right, right in the aisle.

Oh yeah, they're around.

Does it work?

Works. Okay, Alex, thanks for the chat. Very interesting question.

Does your, or has your GTM strategy, the core strategy of what you're doing, has it changed over the last 3, 4, or 5 years with onset of AI? And not the execution, but the core of what your team is doing, is supposed to be doing? I wouldn't, I wouldn't say it's changed. I, from, and I'm gonna caveat this with, I've been at Monday for about a year, so I'm reflecting back to where I think the strategy was. I don't think that AI has drastically changed the strategy. I think I think it's been more about how can we accelerate our execution of the strategy.

Yeah, I mean, I've been at Databricks 18 months. I think the strategy itself, you know, what we're trying to do in the market has been consistent, but how we do it and how can we be smarter in the way we do it has changed.

Yeah, I feel like, and I think, uh, I think the go-to-market strategy changes based on the sales leader or the go-to-market leader more than it does based on the technology that supports it. Understanding AI is disruptive. I feel if it changed our go-to-market strategy at all, it's because of how it's manifesting into our product portfolio more than kind of like how it's used internally for go-to-market. That's a good point. Different, different elements.

Great, great.

Very cool. We have time for 2 more.

Uh-oh.

Hi, quick question on When vendors decide to approach you guys, who are the top 1 or 2 function that decides the AI buying cycle, so to speak, or evaluation whether it's a buy or build?

I can— I mean, I think from— for in our team, it would be a joint decision between myself. I mean, I run strategy and operations, would be myself from the business side and then our CIO. So we would jointly make that decision.

Yeah, I thought for us it would be go-to-market leadership, kind of like usually advocating for those, and then our head of AI and our CFO also in the loop, first to kind of like think about the alternative costs.

Yeah, I mean, similar business lead plus, uh, plus our IT leadership deciding what they, what they're going to approve through into the stack.

It's interesting to hear that Benchling has a head of AI now, so that's, uh, yeah, well, we do.

Uh, one of our founders quit their job and is now our head of AI. Uh, but, uh, yeah, it's, it's, uh, at the end of the day, if you're there's build, buy, or just don't do anything, right? Those are the three kind of like options. And it's up for the go-to-market leadership or sales leadership, field leadership, whatever you guys call it in your org, to advocate and kind of like emphasize the need and potential solutions. And then business technology, AI leads to understand whether this is something we can build internally with less friction and less cost versus buying something either out of the box or customized. I'll be honest, where in most cases we will most likely either pass or buy than build our own just for velocity and speed.

Is it the same in your organization?

Uh, pretty similar.

I mean, we—

I would say we have more of a propensity to build. I mentioned Big Brain, and so we'll go there if we have a real need and see if they've got appetite, um, to take it on. And then if they don't, we'll— we'll go back to looking externally to fill the— to fill the goal.

Yeah, we're a consumption-based business. We find a lot of tools don't gear necessarily off the shelf for consumption, so we tend to build more than we buy.

Thank you. Hello.

Hi, my name is Raj. I'm from Unlead.ai. I love the comment when you mention video workflows. That's amazing. That's why we do it. But the question is, when you decide on a platform and you wanted to do a pilot, is there a number of executives you allocate saying that, okay, 5 people who are going to be using the platform perform, this is the metrics I would need before I'm going to double down or move out. Is there a model you apply before you run a pilot? Just curious. Yeah.

For us, yeah. I mean, we will design, we will design kind of what are the outcome metrics we want, you know, what's the end count that we want the pilot to be. And then, you know, we'll have a, you know, a small team like a tiger team that will remove, review that pilot every couple of weeks, and when we feel like— and we do that for internal projects too, right? When we're rolling out internal AI projects, does it make sense? So we follow a similar pilot POC methodology for both.

Got it.

Yeah, I echo that. We're kind of like in the same boat. We would design some kind of a POC. Usually we have a standardized method of doing that and redefine, like, predefine the metrics that we are looking to see uptick on and then evaluate at the end.

I got same process. You know, there's just so many opportunities out there, so we've got to be structured about assessing them.

Yeah.

Talking about video, if you see videos of me on LinkedIn, that's not really me. That's the AIVC avatar of me.

So for anybody following me on LinkedIn, that's where video is going.

Next question.

[FOREIGN LANGUAGE] Thank you so much for your insights. I was wondering, in your roles, you all have different views of GTM. What would you say at this point are the tasks or the projects that would be non-negotiably human? So even if an AI tool exists for that, you would say, no, I need a person to handle this.

We haven't quite figured out— I mean, going on-site and spending time in our customer's office is an important part of our jobs, so that's a, that's a pretty easy no-brainer. But even backing up from that, right, I think any of our executive conversations, we might use AI to prepare for it and build the material, the content. But, you know, that's developing those one-to-one relationships, especially at the top end of our of our customer segment is still crucially important, and we're not looking at AI to, I just don't see that anytime soon replacing it.

Yeah, I mean, we're very much an enterprise-focused company as well. I think PLG might be something different, or if you're commercial-focused, but I think for us, sales and sales engineering, it would continue to be very much driven by people, just because of the trust and the complexity of thinking through how to deploy things in a customer's environment.

Yeah, I'll say we're a little bit different in the sense that Benchling works in a niche market, which is scientists, lab workers, people with PhDs in biology. A lot of our SDRs, AEs, customer success, implementation managers, most of them have that background. And so AI can help with, again, the administrative or kind of like insight on where things are with customers based off of feedback, but they can't really can replace the demoing, the customers, the scoping of the whatever the customer gaps are that we're trying to solve. Maybe one day, hopefully, but today it's just not at the level of somebody who has master's or a PhD in molecular biology.

So unfortunately we're out of time. This has been a super fascinating topic. Please join me in clapping, you know, for our panelists for the job that they've done here. Amazing. Thank you for that. And thanks, thanks everybody in the audience.

Thank you all.

Thank you. Please take your seats. Our show is about to begin. Thank you. My playbook as an entrepreneur is to get like really frustrated by something and then try to solve that personally. Something shifted where the tools went from helping us do the work to becoming the work. Why do I have like one search box at home and 10 search boxes at work? We should just like fix this. When you open up your laptop and 2030, what do you see? We want to organize your working life for you. Dash is like the best embodiment of solving the same kinds of problems that started Dropbox to begin with. It takes you back to one search box for everything. Literally all of our customers are dealing with the same things. Let's have Dropbox become this kind of lab for distributed work. How do we clean up some of the chaos? How do we design a calmer digital environment? The moments where Dropbox really felt like it was taking on a life of its own. One moment One moment I'm sort of paddling in the ocean alone, the next I'm 100 feet off the ground on this tidal wave. I'm sort of back to being that kid, that little kid who like was programming and just in love with technology.

You can go further than you think. You can go further than you think. Please welcome to the stage CEO and founder Dropbox, Drew Houston.

CEO and founder Saastr, Jason M. Lemkin. Thanks for doing this, Drew.

Thanks for doing this. Yeah.

Thank you.

All right, everybody, let's give it up for Drew Houston.

Woo!

So, pretty fun. Um, quick question. How many folks here use Dropbox? Just quick show of hands.

A few. All right.

How many for more than 10 years have been using Dropbox? Almost the same number, right? Awesome. And how many use it across your org in some fashion? Like, not just for yourself or personally? Certainly fewer. All right, so 10 years. So, I wanna talk about a lot, we wanna talk about AI. AI is the theme. You know, not too many companies more interesting than Dropbox, the amount of scale. So I wanna talk about the future in AI, But I do want to talk a little bit about the present just to give everyone some context. You've talked about like 3 acts of Dropbox, right? And I have a little slide here if folks can see them. I did my own little version. This first one on the left, when's this tweet from? 2017?

2017.

I remember my friggin' jaw dropped when I saw this on Twitter. 'Cause I knew Dropbox was big, right? But no one had ever seen anyone go to a, it was a crazy number, right? I mean, this is in like the 2024, '23 vernacular, I think it's the greatest PLG play of all time, right? Maybe Canva is bigger from— but everyone until AI wanted to build a PLG company that magically sold itself. There was— I mean, there's nothing like this through when, 2017?

Yep, 2017 was when we hit a billion run rate, right?

And it just was it that easy? Did you just come into work every day and it was like, you know, in The Social Network, I don't know if you know Mark Zuckerberg, but there's like the scene where it's a billion, you know, Justin Timberlake, and it was just like that every day, the numbers just went up to a billion. Was it insane? I mean, could you even process that rate of change?

Yeah, I mean, it's sort of, it was a pretty slow takeoff at first, but then once we really got the core product mechanics right, A lot of the early team and I studied a lot of how growth happened in these consumer internet companies and things like Facebook and social media, virality. And it was pretty compatible with the rest of our, like, quant and engineer nature. So we basically built the first versions of Dropbox, and once we felt it was, like, good and reliable, then we shifted our attention from just the product engineering to the growth. And then that is really a game of inches. And the whole freemium model. But it's a game you win in a spreadsheet, because you have this viral loop, this basic viral loop of like, I install Dropbox, I put stuff in there, I share with someone else, that person's new to Dropbox, they install Dropbox, they put stuff in, they share with someone else. And so that viral loop was the same kind of thing that, you know, 7 friends in 10 days, what folks like Facebook really pioneered. But we applied it. In more of a business software context. But it was, when it started working, yeah, in the early days we would have, we would basically print out user numbers and put 'em on the wall. So it'd be like 100,000 users, 200,000, 500,000, million. But then it just started ripping and we ran out of space on the wall. We were like putting like 5 million on the ceiling. But, you know, at that point we were, you know, hundreds of millions of registered users. And when we got to a billion in revenue.

[Speaker] And if you meet, When you meet a founder today that's like super hot AI native, right, that is growing at these type of rates, do you give them any advice that you saw at Phase 2? Is there any, or you just say, go for it, man? Or do you give them, do you help them see across the next plateau? Because they're like, Drew, I went from 1 to 20 this year, I want to go to 100 next year. Do you give them any avuncular advice or not?

Yeah, um, well, it's a process, I think. I think when you're in the early stages, depending on kind of your age as a founder, not your biological age, but just like how long you've been in business, early-stage founders tend to think about the product, and problem, solution, and then more experienced founders start thinking about distribution, like how do I reach a lot of people? Then as you start getting more advanced, or further along, you think more about the business model and the market, And then after that, it's like defensibility and fighting with all the big guys. And then after that, it's like pretty philosophical.

Yeah.

And I triangulated on that myself, because I would have friends who were like kind of going, who were the same, you know, same class as me, same, started the company around the same time. Some folks were a year or two ahead, some folks were 5, 10 years ahead. And they, it was interesting, they all kind of talked about different things. And so you kind of work through that, those maturity levels. So for sure, I think like a lot of AI native founders, first-time founders are very much indexing on the technology, on the problem, on the solution. I think that you have to start there, but you know, as you keep moving up in the playoffs, you have to layer on more and more after that.

Yeah. Okay. So the second act, I, to torture you, I grabbed some stuff out of your public filings so that you get—

Riveting.

Yeah. Riveting.

Here we go. RPU and paying users.

You still do a good job. The next one, like this is, you call it FSS now, right? Classic Dropbox file sharing.

File sync and share.

File sharing share.

File sync and share.

File sync and share. Okay. So it has become a mature cash cow from a financial engineering standpoint, right? 18, I'm squinting, 18 point some odd million paid customers, right? Like insane cash flow from it. A meta question for me, I wonder, why is that the number? 18, like why is it not, in all seriousness, why is it, it's a meta TAM question, because there's, you have a, how many free users are on Drop, I mean it's insane the total number of folks that have ever used it, right? Why does it resolve to precisely 18.x? What's like the big learning from that?

Yeah, I mean I think, I mean I could give you like a mathy answer to that. But I think it's bounded by the size of the market. I think what, if you were to chart what that graph on the left and add the graph on the right, it's like an S-curve.

Yeah.

Right? I say, so when you think about markets, they tend to follow this predictable, maybe I'll do it that way for, I'm not there. But basically you have this period of, you have a period at the bottom where the thing is cooking and it's not, or it's like, It's like you don't really have a lot, and then suddenly you catch product market fit, it really goes up into this exponential, but then at some point you saturate the opportunity or the category matures, and then things flatten out. And then as a tech company, if you're lucky to be around for longer than the first S-curve, you just gotta do it again, and do it again, and do it again. So I think that's a little bit about what you're seeing. We've built a pretty phenomenal first S-curve, and we're very proud of Some of the innovations, or you know, how we've like transplanted some of these ideas from consumer internet, bringing them into business software, having the self-serve viral PLG motion, creating a lot, building a lot of new habits for people. It's obviously, I love seeing folks who have been with us for more than a decade that, you know, I never get tired of seeing things like that. But nothing is forever, so you always have to be getting up on the next one.

But it's like, it's just the math. So Dropbox is probably like 50/50 North America, international, like most folks are, right?

That's right, yeah.

So of the 18, let's say 9 roughly are US. How do you think about— and you've got like 40— you think of your market share as like 40%, right? Everyone thinks about— I think, you talk about it 40. So 40%, you're better at this than me. 40%, 9 million, how many potential customers 'Cause you've got all of them, do you think, in the US, right?

Yeah, so I think the way—

It feels like you have all of them when I look at this chart. Like at least 40% of all of the people that could possibly buy Dropbox are buying it, right?

Yeah, well I think the constraints to growth are twofold. So one is, kinda as you're alluding to, you sort of run out of people on Earth with an internet connection that, have a need to sync files. [FOREIGN LANGUAGE] And so there's not some new continent we haven't discovered yet where people haven't heard of files or Dropbox. And the second is obviously competitive, right? So in any mature category, and certainly in our category, our biggest competition is the free alternative you get bundled within the productivity suite. So we compete against OneDrive, we compete against Google Drive. And for someone who's graduating from college, Today it might be a different story than 10 years ago where, you know, maybe they don't have a lot of files. Maybe a lot of what they do is, you know, they use Google Docs, they use Notion, or, um, you know, it's not that they have no files, but, you know, you have a PDF that you sign here or there, or maybe you use some videos. But, um, you know, tech is always a moving target. The technologies we're all— the tools we're all using keep evolving. And so I think there's also just a I think there's these generational shifts, or these platform shifts in tech, and so I think some of it is you're seeing, you know, Dropbox is on an S— or Dropbox's first business is on an S-curve, but that way of working with files and syncing and shared folders and stuff, to some extent, is on its own S-curve as well.

Yeah.

Okay, so this third phase, just as a student and customer and old days partner— sorry to get up, guys— next phase of Dropbox, I'm quoting, I'm sorry, your investor presentation. We're evolving from traditional file sync and share to AI-powered universal search and content intelligence. So let's talk about that, but it's like a big jump, right? It's, and we'll, so when did you have your AI aha moment sometime in the last 6, 12, 18 months where you're like, the world is changing even more than I thought?

Is it gradual or?

What is your evolution as a CEO on thinking about this as an opportunity and threat?

For me personally, it's probably more like 2016. I think I had a couple. The first one was really in 2016. Maybe not the story you'd expect. So I did, I mean, I grew up as a little kid programming, kind of central casting. I went to MIT, computer science degree, built the first, version of Dropbox. But I never went to grad school, 'cause I got my undergrad, but I immediately jumped into building Dropbox.

You finished college.

I finished college. I did not go to grad school. My parents would not have been down with the early departure program. Although my co-founder dropped out, so we did check that box as a founding team. But no, I was like, "All right, well, as a kid I wanted to be an engineer. I knew I wanted to be a founder of a company." CEO, I sort of backed into it. And as I really like inhabited that role where you're like building an organization and scaling it up, I was like, man, there's a lot of like repetitive, you know, rote tasks here. And like, God, I'm just like dealing with my inbox and I don't know where my time goes. And I could do all these manual things to like tag my calendar, but like, this is what we have. Algorithms and machine learning for. And I just hadn't had a lot of personal exposure to it, so I'm like, all right, well, I'm going to just start using myself as a guinea pig to start building some AI tools to automate the more annoying parts of my job. And so things like class— which of these emails needs a response and which of them— and like, which— where is my time going and how do I tag that with that, or how do I have a computer tag that instead instead of me or my assistant. So, I just wrote these little scripts and like learned a lot of like classical AI and found that what was surprising was like it really worked, even back then. And it was, this was kind of mind-bending for me because like programming started with this premise of like you take a human, the human writes an algorithm, the computer runs it and gives you the output. Machine learning kind of inverted that where it's like you give the machine the output, or you know, the inputs and outputs, It sort of works back to an algorithm, which was really cool. But then of course, as you started, as deep learning and then these language models, they were just language models, I guess, and then they became large language models. As you started going from like '18, '19, it was pretty clear that as computers were learning to like play StarCraft or like generate news article headlines, It was pretty clear even back then that something like this would eventually happen. But then in '22, or, you know, November '22, ChatGPT, that really like threw everything into overdrive.

Was that— I mean, honestly, I— it took— not that ChatGPT wasn't cool when it launched, it took me a while to grok the magnitude that it would do to B2B applications, right? Did you get it in 2022? Was it like, did you drag your management team into a virtual room and yell at them and say, I want to see— I want it— we got to do a hackathon Tomorrow, guys.

We did.

Yeah, yeah, yeah.

We're like, look, this is sort of what the iPhone launch was to mobile, that this is the ChatGPT moment, was sort of the starting gun for the AI era. And so we were quite explicit, like, hey, we need to be an AI-first company, AI-native company. This is happening. We had already seen— and I'll get to our new product Dash. But yeah, it was very clear. I mean, one thing you do learn after a few cycles of this. So Dropbox turned 18 last week. But, you know, one of the few benefits of, like, getting older and having some perspective is you do see, like, okay, everything, like, so much of what's important in tech happens in the few years that follow these platform transitions, or these, like, new eras of computing. And, you know, from the PC, when I, you know, I lived through the PC era growing up there, to the internet, and then SaaS and and mobile and cloud made Dropbox possible, it was very clear that AI was gonna be an even bigger one than that.

Yeah, and you're on the board of Facebook, right?

Yeah.

Did you see anything from AI there that maybe was special or different than sitting on the board of Salesforce or someone else?

What's—

did you see anything earlier from that mass scale of AI there that you brought in or learned from?

Well, I think yes and no. I think The yes part, Meta, I should have said Meta. Yeah, Meta. But yeah, they, first of all, AI and machine learning had, that's like what makes their business, that's what's made their business since, you know, long before GenAI. Yeah. But, you know, a lot of ways it was pretty clear that you could think of like there's a sort of magical black box of machine learning of like matching people's interests to, to commerce, basically, that black box is just leveling up and leveling it up. And then as you have these deep, a lot of the same advancements that deep learning was showing over more classical AI or classical machine learning, those were also gonna send these, make that black box a lot more powerful and really drive the business. So I think huge, very obvious economic opportunity there, but a little different from the, I'd say like the GenAI is sort of qualitatively different, obviously, in a bunch of ways. And I think that was more of an exploration. That wasn't quite as obvious how that would fit in.

I mean, obviously B2C is usually first, sort of prosumer where Dropbox has a lot of it is next and enterprises. But so, you know what I didn't get when Mark at Facebook was just friggin all in on AI, right? And it's, Facebook is such, I didn't get it. Like, I'm like, this doesn't seem an existential, like Google I got, right? I'm just not smart enough. I didn't get the existential threat, but he did. Is it, this is what we're gonna talk a lot about these next two days. Is it an existential threat to B2B if you're not moving quickly now? Do you think it's an existential threat?

Yeah.

Even if you don't totally see it in your vertical SaaS for babysitters, you might not see it, but I think it's existential. Isn't it?

Totally, yeah. And it's a new medium, right? The way— I mean, basically, if you think about it, so much of tech revolves around these, like, eras of computing and these platform shifts. And every time there's a platform shift, basically, you know, either a new startup will have, like, the new platform-native way of doing a thing, and it's a race between the incumbents and the startup to figure that out. Sometimes, you know, sometimes the change benefits the challenger, the startup. So like something like mobile, you know, you saw a lot of like mobile native things that looked quite different from their desktop browser equivalents. I think there's a debate right now of like, you know, to what extent is AI more of an incumbent benefiting mechanism or is it going to be disruptive? I think we're probably in the early innings of that. I think it took, it took a few years after the iPhone was launched to really get the App Store, and then it took a couple more years after that to really get, to go from sort of like, all right, we're gonna take like a, what you saw on your browser and squish it into a phone form factor. I think that was sort of like the first round of apps that you saw, but then truly native expressions of the new medium showed up, like things like Uber or Snap or Instagram, like those could not that really have existed the same way in the prior era. And I think we've seen like the conversational interface, or the chatbot interface, as kind of the first AI-native construct, but I think we're gonna start to see some truly AI-native experiences that couldn't have existed in the prior medium, and I think those will be more disruptive.

[Speaker] All right, I wanna talk about Dash next and what you're doing, but as a segue, like, when you talk to Dropbox customers, customers and the team does, if anything, what are they asking for you to do more of in the age of AI? Like whether they call it an LLM or chat, or are they, are they asking you today, expecting you to do more with their content and files? Are you hearing pull from your base or are you having to drive it yourself? What are you hearing from the, from, from your fields of 18 million customers?

Sure. Yeah. It's, um, So I think, so those 18 million subscribers, and there were half a million, more than half a million businesses, like active paying businesses, and I think a lot of them are asking pretty similar questions. Some of them are about Dropbox, but probably more generally, just like, all right, AI is here, it's, apparently it's everything good and everything terrible all at the same time, How do I make sense of this? And how do I get the benefits of it without having a lot of the obvious threats or challenges? So the benefits we're familiar with, like, okay, yeah, you can— things like ChatGPT, it's sort of like this thing that has a PhD in every subject and can automate a lot of stuff. But there's a lot of new problems where it's like, you know, especially a lot of Dropbox customers skew from— are more creative. So they use— they often are like collaborating around media and video and in creative industries or functions. So you can imagine they have a particularly fraught relationship with AI because if you're a creator, you know, a lot of the way you think about AI is like some other company is like sort of taking all of your creative output and putting it into a model.

It's existential.

Yeah. And sort of automating your job with no financial relationship. Like, that's not great.

AI designed the first draft of everything here.

Yeah.

Our designer then took it. But I could walk you around here this year, all AI. We just— I just went to— I use Rive AI for creative, and I just asked it, design these cool stuff. And it, it can't do 300 DPI perfect, but it's, it's all AI. It's all, it's all— everything designed here is AI.

Yeah, it's crazy. So, you know, super disruptive, um, a lot of new security challenges. So I think on the more on the IT side, you have folks who are like, cool, all right, before I roll out Um, as I deploy AI, how do I do it safely? Because I think a lot of customers start seeing like, wait, my employees are now pasting like PII or like confidential company information into these consumer— into, into things like ChatGPT. And like, that's not good. But how do I wrap my arm around that?

Huge security issues at a practical level, right? And they're asking— the, the bigger customers are all like, are they— because I wonder how— because there is a trade-off in security and all this stuff, Especially, it is a trade-off, right? And are they more worried about it, or do they more want to lean in on the benefits?

I think they're, well, I think they're trying to navigate it, and I think they're looking for partners they can trust. I think part of that's where Dropbox comes in, and I'll come back to that, but, you know, just very basic things. They're like, well, what happens to my data? And I think these are questions all of us have at some level, right? So if I put my stuff in these, especially in these consumer tools, like nobody wants their, like, livelihood or their most important information to be, like, used as feed to train someone else's foundation model or for better ad targeting or things like that. So I think having strong assurances about what's going to happen to my data, is it going to be safe? So safety and security are very top of mind, same way that was with Dropbox 1.0. Yeah. So what we're hearing from customers is like, help me navigate this, right? And help me extend, like, okay, I trust Dropbox and my files. That's great, but as we all know, work— my working life has expanded far beyond that. Like, you know, can you help me with all the stuff that's in my browser or all the other 500 SaaS apps that I'm using? And I sort of have this new problem of, you know, the good news, I have all these new apps and they give me all these new superpowers, but the bad news is kind of the unintended consequences of that where it's like, all right, as I was saying in the little blurb up there, like, some— somewhere along the line, the tools went from helping me do the work to becoming the work. And now work has become— when I look at what's on my screen, it's just like total chaos, right? We have all got 100 tabs open, can't find anything, 10 different search boxes, 10 different inboxes. You know, I think we're all like, you know, feeling the weight of just the overload, especially as we've shifted to these like hybrid working models. We can talk about that later, but we're putting a lot more stress on the system. And, you know, you look at what's on your screen it was never designed to be your workplace. And so what I look at is like, okay, how can we help? You know, there have been multiple evolutions of the Dropbox product, but it's like, all right, in this new world, there's some things we were already working on. Like, yeah, if I were to build Dropbox over again, I probably wouldn't start in the file system, I'd start in the browser. But then as machine intelligence has really been leveling up, then there's this new opportunity for Dropbox to in some sense like solve a lot of the same foundational problems. Like Dropbox is in the beginning, I mean, I was forgetting my thumb drive, that's why I started it. But really it's like, why is it so difficult to find and organize and share and protect my stuff? And back then my stuff was my files and the scattering was kind of across different devices, across different operating systems. But you know, fast forward to 2025 or even like 2022 or '21 when we started working on Dash. It's like the same problems, like how do I find my stuff? Actually much bigger problem 2021, 2025 than it was in 2007. You know, at home I can search all of human knowledge with a Google search. I go to work, I've got 10 different search boxes, they each search 10% of my stuff. It's really annoying. You know, there's these basic things that we've kind of lost. Like when I, you know, when I reboot my computer, the files will are still there, right? But if I restart my browser, like, all of my context is lost, and you kind of have to go reassemble it one link at a time. And so we really thought about these, you know, as I did with Dropbox 1.0, is like, what are the problems hidden in plain sight that no one's talking about but everybody has? And it was sort of like, yeah, the prompt was like, all right, if I were to build Dropbox over again, you know, start in the browser, start with problems like search. Started working on that a few years ago. Also in part, like that was another coding project I did. I was like curious about these embeddings and vector search and things like that. 2019, started building a little test app. I'm like, oh my God, yeah, this, you can like search without keyword matches. It sort of gets what you're trying, you're getting at. I'm like, it searches as fast as I can type. I'm like, why is it everybody on earth is going to have something like this? Yeah. So there are a lot of like ways into it, but You know, again, S-curves. We're at the top of one S-curve. You know, fortunately, things like the AI S-curve is just at the bottom of that inflection. And, you know, huge opportunity for us to both build and ride another one. So we started working on Dash, which started out as universal search, but then added a lot more as GenAI came online. Then it was pretty clear you could also build on that. That's something that naturally added on top. So sort of like Google and ChatGPT but within your own company. Because if you think about it, you know, not only do you have 10 search boxes, but when you ask ChatGPT a question like, hey, you know, what, when does my office lease expire, or what's my agenda for Saastr today, it can't answer because it's not actually connected to your stuff or your context. This was a problem that seemed very obvious, very hard to solve from an engineering perspective. So we've been working on it for a while and And we launched Dash, the business version of Dash, in October, 6 months ago. So we're super excited about it, but it's pretty early.

I want to talk about Dash. I was trying to advance— if someone could help me advance the slide, it's not quite working. I want to show Dash, but we don't really need the slide. But okay, let's do a deep dive on Dash in a second. 2 just quick questions that came up. One, are you coding more because of AI than you used to?

Are you hacking and coding more? Yeah, much more. Yeah. I mean, I don't code during— I wish I could be coding during—

Yeah.

But I am like coding like an 18-year-old again, again, right? You know, starting probably '21, '22, like on my honeymoon I was like prompting, you know, GPT-4, like a different version. My wife loved that 3 years ago. Um, but yeah, I mean, it's again, as an engineer, it's just like, and then seeing like different, the evolution of machine learning, it's just like so amazing where you're able to build. And there were so many like dead ends that I hit in 2016 where it's like, oh yeah, computers can't really understand, can't parse an email, it doesn't understand what's going on. But the large language model basically combined all kinds of like, like 10 innovations in one. So it's incredible what, um, you're able to build with it.

Yeah, one of the— I, I talked with two of the best CTOs I know sometime last year about how they were thinking about AI, and they're like— one of the best ones said to me, if your engineering team doesn't think that they're a kid in a candy store today Fire them. No, he was very, he was like, because you're gonna, like, this is the biggest, if you like to code, this is a git, like, this should be fun for you. Like, you should be getting, he's like, my director of engineering hasn't coded in 2 years, he's coding at night. It's, it's not just to learn, it's he wants to, like, he wants to know. He wants to see all this, so you're doing it again, right? I mean, I mean, you wrote a lot of, I mean, before, I, I don't want to spend too much time in the early days of Dropbox, so I might, I want to do Dash, but You wrote the very first hack for— that was a jaw-dropper. Like, no one could believe that you could hack macOS back in the day and make— I mean, so you're a hacker by nature.

Yeah.

And you're hacking again, at least to some extent.

Yeah. Right? I mean, that's still my happy place. That's like what I love. I mean, I'm like reading AI papers and coding. Um, I, I, some of that, like a mutual friend was asking me like, so what do you do for fun? And I was like, should I really answer honestly? Hopefully, I don't know.

Should. I wouldn't work for someone that has a remotely technical background today that doesn't give that answer.

Yeah.

If it's, if it's they and they're like, "Nah, like I'm sort of hands-off keyboard." Like I would just fuck off and go work somewhere else. This is the wrong time to be hands-off keyboard.

That's amazing.

Alright, so Dash. So, and just for context, like what, so I didn't, so Dash has actually evolved, which you didn't know. 2021 you kick off figuring out how to do universal search against everything. Slack, your own data, Google Drive, whatever you need. And then it radically changes with AI. Like recently it's relaunched. What's the sequence of that evolution? Because I don't know it.

Yeah. So kind of kicking the tires on the concept a couple of years before we started working on it. But 2022, we bought a company called Commandee.

So this was a—

I barely remember that.

Yeah. Was doing— it was doing universal search, but it was local only. So some a lot of limitations, or you could extend a lot further, obviously, if you could do a lot of that in the cloud. And then ChatGPT comes along. I'd been tracking language models for a while, but like, very clear that in addition to conventional Google-style, like, 10 blue links search, you could also build this conversational interface, and that there was this gap in the market of having AI that was like grounded in your stuff or your context. So that— and then, you know, it took some iteration. Initially we thought this was— we were going to do a self-serve thing again, and we built a self-serve version of it, but the friction was quite high. So it was like hard to deliver that, you know, magical experience and deliver good search results. One, if you didn't have all your apps connected, so you're like, "All right, well, you like logging into websites, please log into 50 websites in a row." Wasn't really a great experience. And then there's just engineering constraints where it's like, all right, well, you can link up an app, but then it takes a while to ingest the content and things like that. So we saw that it was hard to get that viral loop going, and then we realized that in most companies, the way you would just sidestep this, you know, Dash, in any kind of large environment, is gonna be something that IT just installs on your laptop for you like anything else. And so you don't even have that onboarding.

[Speaker] IT, IT does it for you.

[Speaker] Yeah.

[Speaker] So whatever onboarding's required, required, it's invisible to the end user, right?

And then as an end user, you just like open up your laptop, and you're like, "OK, cool, I have a Dash search." So we went in that direction, but then, you know, in a lot of ways we've found some new creative engineering and design, other ways to reduce that friction, and we see a big opportunity to build that self-serve version of Dash again. 'Cause it's a little underwhelming, you're like, "OK, here's this cool thing you wanna try," but if you can't actually like go download it and use it—

No, especially for SMEs it's a big deal. If you have to ingest data, right? Or you have to integrate with 50— like, that, that is not a big deal in the enterprise. That is far simpler than deploying Workday over 36 months. It's not— no big deal, right? But for an SMB, even just for our— like, have you guys used our AI for SaaS? Who's used it, right? It's pretty, it's pretty good. You've used it, but it has 19 million words, which is— you can help me, it's not that many bytes, but it still, like, took a full day to ingest 12 years of content. 820 podcasts. Yeah, this isn't too much. And, and the CEO is going to come on stage on Thursday and talk about it. It's epic. But for an SMB, that's a lot of work, right? Right. For a prosumer, stay-at-home, not at the bleeding edge, they're not going to spend a day totally ingesting 12 years of content, are they, into an AI, right? So it's still a challenge today is the point you're making.

It is.

Yeah, haven't totally cracked the code.

And you know, these things sort of look easy, or if you do it right, it looks easy, but it's actually really challenging from an engineering perspective. And with Dropbox 1.0, a lot of similar things. It's like, oh well, all these weird peculiarities of file systems, and like, oh, the Mac handles lowercase letters different than Windows, and blah, blah, blah. So there's a whole new set of equivalence when you're basically— in addition to syncing files now, we're basically syncing every cloud and SaaS app. in a totally permissions-aware way, you know, at Dropbox kind of scale. It's a huge engineering challenge.

But, and then let's talk for a minute about— let's, let's dig in on a few things on it. But what you touched on, the advantage of bigger incumbents that have data versus startups. I mean, you have a trillion files.

Yeah.

So being slightly objective, which is tough, right? How— for this federated— you're doing Federated Search 2.0, which was never interesting before AI. Always, everyone's been trying to solve some sort of federated, like, it's the startup you bought, it's cool, but it never really reached scale until now, right? How big an advantage is those trillion files? Because it's not just Dropbox now with Dash, it's everything, right?

Yeah. So, a bunch of advantages. So certainly the technical infrastructure we've built, um, we, so we're at multi-exabyte scale, so for, that's like billions of gigabytes. Trillion files on Dropbox, right? And we actually started in the public cloud. We relocated to our own infrastructure. So as all these companies were moving to public cloud, we were like the one car going the other direction. But we still to this day, we achieve huge performance and cost advantages by bringing, by tuning everything to our own workload. And we expect the same thing will eventually be true with Dash too. But we're not starting from scratch. So we have a lot of technical infrastructure we're building on. Distribution, so we already have half a million companies already on Dropbox. It's pretty easy conversa— well, it's at least straightforward for people to understand, like, OK, cool, I put all my most important files in Dropbox, got it. OK, I can now organize everything else, great. And by the way, my data is not going to go to any third parties. Awesome. So AI will be safe, great. So it's the distribution and that trust relationship. We have with customers is really valuable. And then, yeah, just our go-to-market motion, like the self-serve viral engine, like that still keeps going, and Dash is gonna be another way we build on top of that. And I could keep going on about, you know, there's a reason I'm building this within the context of Dropbox versus like starting a new AI growthy startup, which, you know, there's some days you're like, man, this is like, trying to get all this wired up is a lot of work, but But no, there's huge advantages to being established, and like a trust advantage, and the fact that we're platform agnostic, right? We don't favor any one platform. It's gonna be just as good an experience on Microsoft's ecosystem as Google, as Apple, and everyone else. So those are all big advantages.

[Speaker] Just a question, 'cause I'm only so smart and a little ignorant. When Brian Halligan from HubSpot talks a lot about this, and we did a catch-up with him a long ago, he's like, 'cause they're all in just like you, right, at HubSpot, right? And he's like, just don't compete with us, do it, just don't underestimate how important the fact we have all this structured data is, right? It's great that you can chat, right? Dropbox, like, and again, you're much better, it's sort of semi-structured, right? It's tagged. I've got 12 years, but it's not, but do you have the same advantage versus incumbents that a CRM or an ERP has that has literally built thousands of lead contacts, opportunities, and such, right? So how much of an advantage does it give you? I just don't know. I just don't know.

Yeah, I think it's more in like the distribution and brands.

That's huge.

And technical infrastructure and things like that. I mean, there is a data flywheel to some extent too, though, as you'd imagine. We're, you know, we start with privacy, so we're not like training any models on the contents here at Dropbox or anything. So, so we don't, you know, we're very careful about not crossing lines like that. But yeah, I mean, certainly that, it's just, you know, we had different sort of, or we always knew we needed to build like another S-curve. And so like going multi-product for us took a little longer than I would have wanted. Yeah. We had some false starts there, but it was really important to me that, and Dash ended up being the ultimate product, but I'm like, we really need something that's like a natural evolution of what we do. Do. You know, sometimes you can have an Amazon that starts by selling you books and music and then like plot twist goes into like cloud compute, like selling compute as a service. But for us, when we went too far afield, like it was a real challenge to get people to think about the company that differently. And so I wanted something that, where it would sort of stack on top of what we already built.

All right, so one just super obvious question on all of Dash, which you've answered, I think, before and wrestled with. So Dash is cool, I can integrate Dropbox and all my other systems. It's not quite a 1-second process, but it's world-class, it's getting the best IT can do it. But competitors can integrate Dropbox into their Dash Prime, or what's Dash backwards? Hasad? Give me an acronym. I'm just having some fun, right?

Sure.

But you're not moving to a closed model. Right? You're gonna let people OAuth and others and get their content that they're storing and paying, right? So how do you win in this hyper-agile world where people agree with you, but they have the same vision? Right? It's not, not just even other, like, but a lot of people want to do what you want to do with Dash, right? And, but you also have to support them. It's, it's, there's a, you have to compete with it and can cooperate with their efforts, right? So how do you win in this I'm a super fan, but how do you win in this— when everyone's trying to do some— at least a little piece of what you're trying to do, right?

Yeah. [Speaker:JASON M. LEMKIN] Um, so a couple parts to this. I mean, first is just the basic customer problem, especially with our customer segment, right? When I look over someone's shoulder in a Starbucks, they don't have a search box that works, right? And we certainly have— we're not the first company to enter enterprise search. The idea's been around for a long time. There are companies like Lean and others, but they focus on like very large enterprise customers. And so, when I— most of the time when I talk to a customer, they have no options. So it's like, right now it's white space. And frankly, even Dash doesn't have like the self-serve. We're building towards that, we don't have it yet. But we see that as like a natural place where Dropbox can be really strong. Where, you know, frankly, you know, upon launching Dash, we'll be the only game in town. Despite AI generally being this very competitive landscape. So I think, you know, one of the lessons I've learned over fighting with the Microsofts and Googles of the world for most of my adult life is you do have to like focus, be like, okay, what market are we really gonna be in? How are we gonna win? What are our structural advantages? Let's play the chess game ahead of a couple moves. Like, what are they likely to do to respond? And so on. So, but it starts with a customer problem. Like, people do not have either a search box that works, and they're also not using AI that's connected to their stuff. Right, they're using these consumer tools that, and then the integrations, like they might say like, oh yeah, you get stuff from your Google Drive, get stuff from this, get stuff from that. But like, it's really not comprehensive. Like, these are very like surface-level integrations, whereas Dash builds like a full index of like all your companies, of your contacts, your company's contacts, and that is super necessary. Necessary to make the— to have the intelligence be intelligent, you need to give it the context, and that's what we do. Now, so from a customer perspective, I think we're going after a lot of white space. This is gonna be a universal knowledge worker problem. It's really— it's harder than it looks to do, um, and so we're putting a lot of— we've been putting a lot of resources here, and I'm pretty optimistic about that. Um, and it reminds me a lot of Dropbox 1.0. People didn't think of Dropbox as a file syncing thing. We weren't the first storage thing, but, you know, we tried to be the first We first want to get it right and get it to scale. But you were also talking about developers, right? And startups. You know, it'd be one thing if the intelligence was kind of isolated within the Dash UI, but how many startups, probably even in this room, have like a business idea that starts with like, well, we're going to build this thing, and step 1 to build that thing is we're going to need to build 50 connectors to your Slack, to your Google Docs, to your to your Salesforce, to your blah blah blah, to like make this intelligent, you know, SDR. Like there's a founder running, who I was talking to, he's like, "Hey, we're building an AI SDR, we'd love to bring in the context from Dash, can you do that?" And the answer will be yes, right? So we also want to empower startups and builders to be able to benefit from a lot of the context that we're assembling in Dash, because we've put hundreds of millions, we can put hundreds of millions of dollars and have of R&D investment into actually doing that properly and securely and reliably and scalably and all the things you want. And then through MCP, which is this Model Context Protocol, and, you know, other API integrations, we can bring the intelligence we're building in Dash to other surfaces. We, we don't have that this second, but— or we're just in the process of like ramping that up. But absolutely want to bring the intelligence from Dash not just into your browser or onto your your computer, but into all these other intelligent experiences that you're using.

Just one— we briefly talked about it backstage, but just maybe just an interesting GTM question. For me, when it gets there, when I log into Dropbox— and I use Dropbox all the time. I never log in. But when I go to the website and all of a sudden it's Dash, and it's not just Dropbox, everything's there, great. I'm all in as an existing customer. And especially if it's included in whatever we're paying, I don't even care Dropbox is not so cheap, but even better, I'm a super fan, my NPS goes up, my retention goes up. But you talked about getting customers that aren't in the 40— the 18 million. They like— I know it's 500,000, 18 million. I'm gonna use it, like, whether I'm gonna pay for it or exactly, like, because I'm already in, right? But you want to unlock the rest of the world, right? And so what— I know it's going to be great, that's— it's going to be S tier, right? But why will they look to Dropbox, the other Beyond the 18, 19 point something million pains, you know, at least seats you have, why will they come to you for this federated product?

So, uh, couple reasons. So as you pointed out, with our existing customers, we have this huge home field advantage, a relationship. We already have your files, you know. Again, that is one conversation. But, you know, even at that kind of scale, this kind of scale where we are, you know, 18 million subscribers When you look at, there's like several hundred million Productivity Suite subscriptions in the world. You know, what that's really saying is like, and you look at the evolution that we've had from being on desktop apps in our file system to being in our browsers, you know, it's something like 95% of the TAM are gonna be people for whatever reason aren't doing file syncing in Dropbox. And so it's like super important, Dash, you actually don't need files in Dropbox. It works great if you do, but it's a separate product for that reason, because we wanna open up the whole TAM. And frankly, we'll just as happily organize your stuff in OneDrive or Google Drive or anywhere else, just like we would the content we have in Dropbox. And this is part of, we're trying, there's a lot of reasons for this. I mean, we wanna disrupt ourselves. You know, tech is always a moving target. And then we look a lot at the great, you know, we're going through, our generational transition, you know, navigating that, that switch in eras of computing. And, you know, that's a hard thing to do. And I think it's something you only really get to do if your company's been around for, you know, 10 years or more. But at some point, your company absolutely will reach the top of one S-curve, have to start another. And when I've studied, you know, as you'd imagine, I've studied all the companies that have done that successfully, you know, Adobe going from selling you Photoshop in a box, to SaaS and Creative Cloud. I think Netflix is a good example I think about and talk about a lot, right? So they were very early with content and digital content. They started out by mailing you DVDs, and then they realized that the future was gonna be streaming, and they had to navigate that transition, right? So they had this DVD mailing business, which was, you know, a pretty reasonable scale, but also capping out, and not the future. But the best, turned out the best thing they could do for their DVD mailing business was actually add streaming, not like put, make a better envelope to put the DVD in or something, right? Because that, being able to like bring your customers along that transition seamlessly, like it just showed up as Netflix on my credit card statement the entire time. Yep. But that audience helped them bootstrap a victory in the streaming wars, right, which were equally competitive, maybe different competitors. But most importantly, unlocked a TAM that was like 10 or 50 times bigger. And you know, it's not just Netflix, newspapers had to go through this, there's a lot of examples of how do you navigate that S-curve, and I think that's something, you know, as a founder in the double-digit year kind of cohort, you have to learn from these kinds of transitions and design them intentionally. And so, I mean, there's some differences, like files are not not going away, right? And we have a— there's a reason those 18 million customers are on Dropbox and not somewhere else. We do a lot, especially if you're in the, as I was saying, the creative community, if you deal with big files, if you do a lot of external sharing, Dropbox works equally well with every platform. So there's a lot of advantages we have in FSS too. I don't mean to say that it's done, and it's certainly not going away. But you have to both address— it's one story for our existing customers, but we also want to unlock a reason for all the people who didn't raise their hands about using Dropbox, give them a reason to use Dropbox and Dash.

All right, last one, 'cause we're just about out of time. There's other stuff we'll talk about at a sequel sometime, go-to-market and teams. There's something that you said that I've been telling every, almost every founder I talk to, which is, unless you're growing at insane rates today, you have to refound your company day in and day out. You have to go back to first principles. And I literally— we're not all at Dropboxes. What's your, what's your ARR again? I get—

it's 2.5 billion.

Okay, 2.5 billion. But I say, listen, protect your base, protect your— but you've got to sit down and say, what company would I build in 2025 if I— and this is exactly what you did, right? So just one tactical question: how much of your time do you sort of spend on Dash and AI versus the existing 2.5 billion of AR, because it is a cognitive challenge, isn't it?

Yeah, yeah. Well, first on, you know, how to navigate these things. I mean, these inflection points in tech, or these changes, there's a lot of history here that I think every founder should learn. If, you know, if— and frankly, a lot of what I learned was from reading Only the Paranoid Survive, which is by Andy Grove. He was one of the early engineers at Intel. He became CEO. And Intel is navigating this transition. They actually, we know them as— Again today. Yeah, so again, it never stops. So I think there's like understanding the history and the theory and the grounding in that. And then, yeah, the answer is I spend a lot of time on Dash, but I still, I spend a lot of time in, I am quite scattered across many parts of the business. But then also, you know, on the core business and then also like building that bridge between the products, right? Because we want that, we don't want to build two, it would be ironic if our contribution was to fracture the content experience even more. 'Cause you shouldn't have one UI for your files and a completely different UI if it's a link. We want to have this coherent, cohesive experience. But I also still spend time thinking about, okay, there's still a lot I would do to keep disrupting our stuff. Ourselves. And sometimes you think of it sort of negatively like that, or if I wanted to like attack us, I would do one thing. But another is just like, there's so much room for reinvention in general, and we sort of become blind to what we see because we're used to it. But, you know, digital music had many, many chapters of evolution, right? From like literally ripping CDs into MP3s that were in files and folders to then iTunes is like, this is dumb, you should have like a music catalog, there should be artists and album and playlists. And then Spotify is like, wait, why am I paying for songs one at a time? This is medieval, I should have, you know, everything like an infinity catalog. And by the way, music should not— should, um, it should be multiplayer, right? I should be able to have shared playlists, I should know what my friends are listening to, it should be intelligent, right? I shouldn't have to add songs to a playlist one at a time, I should be able to have I should have smart playlists, I should have like an AI DJ, I should have stuff in between like these auto-generated playlists. So, you know, there's like this drumbeat of innovation that's happened there, but then when I look at the content, and that's like the content experience at home, right? And everything works this way, Netflix, Spotify, YouTube, they don't like show you like, here's 15 million movies starting with A, right? It's like this auto-curated intelligent learning experience. Then I go to work and it's like, where is that? Right? You know, why do I have like a file browser on one side of my screen that's like—

hasn't—

literally hasn't changed in 40 years, and why do I have a web browser on the other side of my screen that has so many tabs open I can't even see text or titles anymore? Yep. Like either this is it and we've cracked it and this is what we'll be doing in another 40 years, or there's a better way to do it. So I think you have— and then so then it's like, alright, well, let's work back from that like ideal solution. Let's figure out what we can build today, you know, sequence it. But, you know, fortunately for every problem that technology solves, it creates another one. That's what makes it a great business. But, but no, it's all, you know, there's always opportunity to reinvent. Your customers are never like, we're good. There's always things you can be doing better.

All right, last question because we're over, and then we will break. For folks here, because it's common, believe it or not, it's common even at small startups, this not tiny ones, but where pockets of your team are resistant to AI change. They don't want to use the tools for whatever reasons. I hear it in engineering, I hear it in sales, I hear it in marketing. How many employees do you have at Dropbox?

2,500.

Okay, so you've thought about this issue.

Yeah.

What's your best advice in a, in a, in a, you know, to help folks see, adopt, and change embrace it and move more quickly? What's your, what's your best advice thinking about 2,500 folks?

Well, first I think, you know, I think most tech companies, at least the CEOs I talk to, like, have had this sort of surprisingly low adoption organically, right?

Surprisingly low, right?

Which is a problem. So I think you have to do these, you know, gently but firmly, but you have to have a mandate. Be like, look, you know, the future is It's going to be here. We're here. We need to get there. But I think there's a couple of things you have to disentangle. The people who are resisting it are often— they're not stupid. So I think you have to be responsive and empathize with their concerns. They're like, look, these things look more efficient. It's a bit of a mirage. You're sort of trading one set of problems for another set of problems. They hallucinate. They may They make mistakes, blah, blah, blah. So like, they're not dumb for like pointing that out. And so I think you have to be like, yes, that is true. However, it is also true that the water lines, these things are like improving. They're compounding at a massive rate, and that's gonna go from right to wrong real fast. Second is, look, I think empathize with them on a more human level. Like, sure, like, imagine talking to a photographer and telling them, hey, you gotta leave the darkroom and now go sit in front of a computer and use Photoshop, right? Go from there where you're really comfortable, you're like a craft expert, and now you're like totally a novice and embarrassed and like starting over again. You know, on the one hand, that's like good for you to stretch yourself, but that you can just at least recognize like this is uncomfortable, right? And it's a new skill, you know. And you have to say like, look, you know, I'm not your parent and I'm not going to tell you what's good Good for you. However, I don't feel like I'm doing you a lot of favors in a world where, like, if you're an engineer saying I'm not going to use coding tools, it's like a finance person saying, "Hey, I'm just not going to use Excel." It's like, "Okay, that's your prerogative, but I don't think any other company that's going to hire you is going to think about this any differently than we will." And so, and then by the way, often the people who are most religious about it have the least hands-on experience. So it's like, please educate yourself. And like, let's, you know, we can debate how good or bad these things are, but everybody's moving up the learning curve together. And, you know, we're not doing our jobs if we're not getting you ready for a future that's coming super fast.

Yep.

That's good advice. All right, everybody, let's give it up for Drew. This was great.

Thank you.

Thank you. Thanks for coming here.