David Sacks & Evan Owen - Glue: The AI Solution to Workplace Communication

Welcome to the Casey Adams Show. I am your host, Casey Adams, and over the last six and a half years, I've interviewed hundreds of world class entrepreneurs and founders. On today's episode, I sat down with the one and only Evan Owen and David Sacks, the co founders of Glue AI. Glue AI is a new work chat app for AI assisted conversations. On today's podcast, we did something different that we've never done before, which is an entire product demo live on the podcast. So if you are listening, make sure you head over to Spotify or to YouTube to see what it is that we're talking about throughout this episode. Last but not least, if you're not familiar with David Sacks, he was the former COO of PayPal, the co founder of Yammer, which got acquired for just over a billion dollars back in 2012, and of course the co host of the all in podcast. This is a unique story where Glue was incubated inside of his venture fund, Craft Ventures, which with Evan Owen, and they've been working on this product for the last handful of years. And finally they are accepting beta users. And today we're here to talk about it. It's something that not only I'm a huge believer in, but seeing how they're thinking about how AI can impact enterprise communication is just incredible. So if you're just watching this for the first time, thank you so much. First of all, please take a moment to subscribe, share, share this episode with a friend if you've got any value from it, and enjoy today's podcast with Evan Owen and David Sacks. All right, welcome back to the Casey Adams Show. Today we are joined by Evan Owen and David Sachs, the co founders of Glue AI. Guys, thank you so much for being here today.

Yeah, thank you.

Absolutely. So, David, I know I had you back on my podcast. I think it was 2021 on the call in app. So pleasure to have you back here today. I saw that not only what you guys are launching at Glue AI and what you brought to the market, but for people that may not know what Glue I is, you guys recently publicly announced this. You've been working on this product and company now for a couple years. Would love to get a high level before we really dive into the story of how we got to where we are today.

Sure. Glue is basically work chat for the AI era. I've been interested in the space for a long time. I created a company called Yammer way back in 2008 that was one of the first enterprise social apps to basically try to tackle this problem. Of workplace collaboration. We ended up selling that company to Microsoft back all the way in 2012. And that product was really based around Feeds. It was sort of really kind of well ensconced in the whole social networking era. Then around 2014 you had the rise of Slack and channel or IRC based communication. And now we're obviously in the era of AI. And if you've used ChatGPT, you know that it's more thread based communication where you create a thread to talk to the AI. Each thread has its own name and it's sort of topic based. But obviously ChatGPT is not multiplayer and it's not a fully featured chat app. So we've come to market with a new product that I guess you can call it kind of a mashup between Slack and ChatGPT. It is threaded like ChatGPT, but is a fully featured enterprise chat app. And this thing I've been working on with Evan for a few years now, and I'll let him tell his story in a second, but basically Evan joined craft as an EIR, I think way back in, was it 2020 or 2021? And he had tackled this problem creating. Creating another company which was. It was what you guys call that.

I don't think Zynk.

Yeah.

Which got acquired by ServiceMax. But anyway, why don't you tell that story?

Sure, yeah, go ahead, Evan.

Yeah, so you know, previous to David reaching out to me, I had been working on more communication for a little while. Zynk was a mob messaging app for field workers. This was right around the time that Slack was starting to take off and so spent a lot of time thinking about how people communicate, what you're trying to do when you go to communicate at work. And then when I started talking to David, I realized he and I had both been thinking about a lot of the same things and there's still a lot of innovation to be done in the work communication space. So that's kind of where this all started.

Yeah, no, it's so fascinating too. And I love that when you say eir, David, that means entrepreneur in residence. Is that something that you guys typically will do at Kraft, right? Like having a founder inside the company to then partner with and bring this product to market, Even outside of just investing into a company or, you know, what was the distinction there when it comes to this product with Evan and, you know, bringing it to what it is today?

Well, I guess what's different is we kind of incubated it ourselves inside of Craft and it's not the only time we've done it. But it's rare, you know, we don't do that many. Calling was an example that we had, we had done that before. So, you know, once in a while we'll get an idea that we feel like should be pursued and no one's exactly doing it the way we would do it. So we, you know, incubate the idea. But again, this is a case where, you know, I've been interested in this space for a long time, over a decade. And you know, Yammer got acquired by Microsoft so quickly that, I mean, it was basically a four year run from starting the company to getting acquired that I felt like I had all these ideas that I never got a chance to explore. And I think Evan felt the same way. And so we teamed up to collaborate on this and then the whole AI revolution happened in the midst of that and we realized that the direction we were going was kind of perfect for AI and so we leaned into that even more.

Yeah, no, that's, that's so powerful. And just for the listener today here, in a bit we'll be doing, we'll be jumping into the screen sharing and really diving into glue. And Evan will be doing a great walkthrough. But Evan, I want to jump to you for a second because I know and even David, with you having the all in POD and talking about this story with Evan, other shows people are very familiar maybe with Yammer and what your background is there. But Evan, the company Zynq, that was in this sphere of enterprise communication. Talk to us about your background there. And I know David said how you still felt like you had something to do there, bring us that perspective just before we really spend some time diving into the product.

Sure, right. So I started at Zynq as an engineer. I have a technical background. I was one of the first engineers on the Zynq team and I was able to work on the product cross platform. So doing some iOS development, web development, back end, eventually serving as the VP of engineering for Zynq later in the company's history. And honestly my passion has been in the technical side of software for a while. But as I spent time on the product, I became, became more and more interested in just what users are experiencing, like what makes this product easy, what makes this product fun to use. And it just slowly started becoming a passion of mine and I realized that I really wanted to be involved in the founding stages of something and really have a lot of impact on the product. And so that's one of the reasons Why? I was really excited about partnering with David on this and just working on something that I hope, I think will bring a lot of value to people and will improve people's everyday lives.

Absolutely. And I know I've heard you guys talk about this before, but this product isn't just something that you guys spun up an mvp. You put it out to the market, which we'll get to. But you guys have been working on this, as you said, David, for the last two, three years now. And it really brought it to a place where you wanted to put something out into the market that not only was polished, but was able to work with these larger enterprises to get them onboarded in a way that's much more efficient. And that's not just, hey, let's put something out there and you know, workflow, but really have something that's professional and adequate to make, you know, these big transitions from a company like Slack that, you know, is a competitor in this space. Is that accurate?

Yeah, absolutely. No, I mean, we, we, we wanted to come to market with a product that was not just an mvp, that was actually highly polished. In fact, we've been using it inside of Craft Ventures for over a year now for all of our internal needs. And we didn't want to launch it publicly to other teams until we felt like we had it really dialed in. And the reason is just because although we think that there's major issues with Slack, people still like Slack. It's a good product. And so there's no point going to market with a product that doesn't hit a certain bar. We have to be a lot better than Slack. And I think we finally got to the place where we felt like we were and so we were ready to launch.

Love that. And just before we jump into screen sharing here, Evan, David, I want to ask you, just reflecting and even kind of thinking about this in the moment here, reflecting on your past with Yammer and you said that was the four year run from 2008 to 2012 and that was a massive success. When you think about what went right at Yammer and what you hope to bring to the table in terms of this market to glue AI, what are some of the parallels that you still see even though a decade has passed and you're bringing this technology that's in the similar market, what do you think needs to go right and what do you think are the biggest hurdles, having done it before?

Well, I think a big similarity is that back in 2008 we were at the beginning of a new era or paradigm inside enterprise software. And specifically what was happening was that people or developers were starting to use consumer software concepts in enterprise software. The first thing that happened in the early 2000s was just the move to the cloud. I mean you got to remember that in the 1990s, enterprise software, just business software was on premise. You know, it was like on a disk or on a server that you ran behind your firewall. The whole idea that software could be delivered to businesses as a website was pretty new. I think Salesforce got started with that concept in 99 or 2000, roughly got it. And so even by 2008 there was a lot of skepticism about whether enterprises could trust cloud based software. We were still working through those issues. And then you had the whole social networking revolution happening around 2004, 2005. It was starting to get really big around 2006 through 2008. And our view was that okay, we fought this cloud battle. We're totally bought into the whole idea of multi tenant SaaS being the future of business software. But once we have that, why can't we design business software like it was consumer software? Why can't we use a B testing? Why can't we use virality? Why can't we just go viral in an organization through employee adoption? Let the end user just adopt it and start spreading it as opposed, as opposed to having to knock on the CIO's door and like sell it top down. So that was the beginning of the whole bottom up or product led movement in enterprise software. And so we were able to combine a couple of these concepts. The idea of taking social networking from the consumer space, bringing it inside the enterprise as a collaboration tool with a bunch of viral tactics that we had learned going all the way back to our PayPal days. And that's what enabled the product to kind of explode and grow very quickly. I think we're in a different era now where everyone has already accepted and embraced all those lessons and all those challenges we were kind of making to the established order back in those days. I mean everyone now knows that business software by and large should be in the cloud, that you want to use the same software best practices you use to design consumer software that you can use, product led growth, that you can go viral, all that kind of stuff. But I think the similarity now is that we're at the cusp of this AI era. So whenever you have a new huge tech wave, it creates a lot of opportunity because a lot of things get disrupted. So that is a thing that's just happened over the last 18 months, I think, is tremendously useful for us is because it's allowed us to lean into AI as a major differentiator for this product.

Absolutely. And I love what you've said of just you're bringing the chat aspect of AI to the workspace and them living independently is one thing, but it can become the superpower if brought together. So on that note, Evan, I would love to take some time here to really dive into the product. And for everyone that may be listening on Spotify or Apple podcasts, make sure you go over to the YouTube video or if you're watching, if you're listening on Spotify, pull up the video for a second because we're going to be diving into Glue a screen share and really walking through the product so that you can see it, feel it, experience it, and we'll be diving into questions and just seeing the magic of what they've created here. So, Evan, take it away.

Yeah, perfect. Thank you. So, yeah, I'm just going to share a screen and show you how we use glue in our own team to build glue, which I think is a great example. And I'll just walk you through some of the basics here and kind of show what's the same and what's different from what you might be used to. So, first of all, just want to start here in the big space on the right. You can see that in the core conversation we have a lot of familiar concepts. Like David said, one of our early tenants was, let's not try to just reinvent things that are working well and that people love about work chat. And so you have a lot of the, the things you expect to have here. Of course, text, it's real time. You can react to messages, you can attachments, images, video, audio, all that. You can send code, snippets if you're a dev team, link previews, on and on. So that's a lot of similarity. Also, you'll just notice in the ui, we have some very similar UI layouts where you can kind of see all of the conversations that are for you on the left and select one and then see the conversation on the right. That works really well. So we're really embracing that.

Yeah, I love it too. It's like the familiar concepts of, you know, when people are seeing what they see here.

Right.

If people are still listening. Right. Like the similarities to Slack is ever present, but the ability to lean into what you guys are wanting the users to really focus on. This idea of threads, talk to me about that. Like, what is the true differentiating factor around threads and why people and users, you know, should be excited about what you guys are creating here with threads inside of glue. Right?

Well, I think what people discovered after a couple years of Slack usage is that really, to keep things organized, you kind of have to do the conversation in a thread. And so a lot of companies even enforce this behavior of, okay, you know, have your conversation in the thread, don't keep it in the. In the channel, where everything gets mixed together. Well, I think there's a good reason for that, right? And it's that when you have that focused context, you can really easily know what you're talking about, know where to send that message, know where to go back and find it later. And it's the same reason why ChatGPT has, you know, these topical threads so that the AI can have that focus context. And. And that's really what we saw sort of coming together there, and we just decided to embrace it. Hey, what. What would a communication app look like if the thread was the first thing, you know, the primary thing that you interacted with? And so that's what we've done here. So here.

And just. Just to add to that. So, you know, obviously Slack has threads that are created when you reply to something in a channel, but all the threads are within a specific channel. And in order to see any particular thread, you have to be a member of that channel, you have to join it. Well, on a long enough time horizon, everybody ends up in every channel because there's one piece of content that you want to see, there's one thread you want to see, and that creates an incredible amount of noise. And that really is the problem with Slack. And what we heard from users is that they really wanted the threads to be basically jailbreaked from the channels. So we have this much more flexible concept of groups, where if you want to just post in the group chat, you can. But you can also address a thread to specific people or to multiple groups. And of course, if you decide on a thread later that you want to add somebody, you can just add them to that thread. They don't have to then go join the group. So it's a much more flexible addressing model. And that way you can just be in the groups that you really care about and trust that other people will add you to the threads that you need to see.

Absolutely. No, that's. That's incredible. And something that I want to bring up here that you guys were speaking about on this Week in Startups with Jason, when it comes to the onboarding of companies at this stage. And I'd love to dive into what you guys have learned so far from opening the product and taking on businesses. Because you said something that really stuck out to me, which was this superhuman approach. And we'll definitely go back to diving into the product, Evan, but the superhuman approach of taking companies onboarding, making sure that they invite their first five or ten employees or people on their team to integrating their apps, um, why is that so important with a product like Glue? And what have you learned so far from that slow, you know, user onboarding experience?

Well, I think that, you know, in the old days, what you probably would have done was just launch and then everyone signs up at once and then you get a bunch of feedback on whatever issues that you didn't anticipate. There could be bugs or there could be paper cut issues or little usability things, features that you realize you're missing. And I think what superhuman did that was actually turned out to be a pretty good approach, is that they kind of did this white glove onboarding for every single user. And so they explained to every user kind of the power user features and like how to get the most value and they help get set up. And so I think that's the stage we're at right now is we're using this kind of white glove onboarding to pace the number of people who are signing up so that as we catch these little issues, we're able to fix them.

Yeah.

And then the other thing, and we're getting a bunch of learnings out of that. And then the benefit for the user is they get set up properly. And I mean, after doing this for some period of time, I think we'll probably just throw the gates open and let everyone sign up and it'll be completely self service. But until we learn exactly what that ideal self service flow is or process is, there's a lot of benefit to us. And just kind of doing this very deliberately.

Absolutely. And I think on that note too, Evan, I would love to have you bring up the screen for those viewing just to showcase some examples of the power of glue.

Right.

I know. David, you talk about this concept of AI being dependent on prompts right now and how things will move to being promptless. I would love to showcase the people viewing today, like, hey, you're signed up for glue. What are some of the key use cases that people should get familiar with once they are onboarded? And then what are some things that they may not be familiar with that you believe will become more important, maybe in terms of the use case, whether with big teams, small teams or anything in between.

Yeah. So I think a great way to do that is just to walk you through this idea of starting a conversation with Glue AI. Glue AI is our AI colleague that works alongside you in your workspace. And then we can kind of, you know, grow from there and add some more folks to the thread and I can sort of show you the whole value flow here. So this is our composer. This is how you start a thread. One of the things we've done is actually bake a lot of AI improvements directly into the experience. So as you type a message, for example, we'll have AI generate a subject for you so you don't really have to think about that. I usually choose that suggested subject nine times out of ten. It's pretty good. The other thing we've done is suggested recipients. So very often you can just click add on the groups or the users that it's suggesting and it gives you what you want. So let me actually just add a different group here, just for sake of argument. We'll say put this to development group. Actually, I want to show you what happens when we talk directly to Glue AI. That's right, but that's the composer. So what's great about talking directly to Glue AI is it basically brings a chatgpt like experience right into your work chat app. And so if you're just starting a thread directly with Glue AI, you can do most of what you can do with ChatGPT, ask questions, brainstorm, get it to write some things for you. But we've also given it full access to all of your glue conversations. So here's an example of something we can do. We just love some updates on current projects. And so we'll just start this thread and Glue AI will go out and it will search. And because I've mentioned the group, it'll search in threads in that group. You could also mention a thread, you could mention people. It really has full access to everything that you have access to. It's your personal assistant.

Yeah, that's so incredibly powerful. I mean, especially as you know, it's the use case of as you use glue for longer and longer and the knowledge and data can build within that is beyond powerful. And I'm curious to know, when it comes to small teams, large teams, what have you guys been seeing in terms of feedback so far post launch of the different use cases? Of course, whether that's for managers, whether that's for the CEO or just people that are looking to backtrack and remember what they said or the key projects. Talk to us about some of the use cases. And maybe too, if you want to showcase some of the different ways that people can dive into that.

Sure. I think, David, you probably have a couple that you've really used as a lot as well. But here's one thing that I really like, especially if you have a large team, sometimes you don't know who has the answer to that question that you've got. It's maybe an obscure thing or maybe who's an expert on a particular area, but you can actually just ask who knows the most about iOS development.

Right.

That's so great. Right. And so of course Glue AI is going to search for a particular topic and we can do semantic and full text search. And then Glue understands who is saying what and can very easily deduce who you should be talking to.

And that is for people that are curious. It's the ability to backtrack into conversations happening within the entire organization to then pick up when iOS is talked about through these models. Is that correct?

Exactly. So again, we're giving the AI the full context of the conversations that you have access to, and you can scope it if you want, but then we just provide that to the model. Also using rag, using vector search to make sure that the most relevant content is given to the model and then the model is able to figure it out. And we're using the latest models by default. We use OpenAI's GPT4 model here, but you can also choose to use CLAUDE if you prefer. And we just want to continue providing the best model capability as the models advance.

Just point out one small feature there. You may notice at the bottom there's a line that says source.

Yes.

And it will actually link to the threads that it found that it used to create that answer. And you can just click on one of them and then we have this neat split screen view. So you can just go directly to that.

That's so incredible.

Yeah, so it's really neat to be able to do that. I mean, it's showing its references or citations.

Yeah.

So one of the other things that we've had users really love to do is take other content that they have and it might be, say, a bunch of documentation or part of your wiki. And we're building lots of integrations that will make it really easy to pull this content into glue. But then they'll put that in glue and then they'll let their team ask questions against that content as well. Right. So you can imagine that you could be searching and finding the most up to date information from all your chats, but then also correlating it with your documentation or, you know, the sort of corporate policies or whatever is important to you that you want to give access to. Your team. Glue AI will be able to have access to all of that and answer questions that way.

Yeah, I don't know if you saw the episode the all in pod where I demoed glue. What I did is I put all the transcripts from our previous episodes into glue and so I could just ask questions like when have we discussed Google AI? And it will give me the episodes that we've discussed it and I can say, give me a summary of what we said and what did each person say? And we even said if this topic comes up again, can you predict which bestie will be pro or anti Google? Stuff like that. And it did an amazing job. We asked it about describing the besties personalities and it was actually able to give really accurate descriptions of each bestie based on all their contributions to the podcast and how we kind of work together as co hosts. I mean, it was really pretty incredible. But now because we did that, we have all that content at our fingertips when, when all in is just working as a team behind the scenes, we can just ping Glue AI and it will know a bunch of stuff about past episodes so no one has to go look anything up. Another example that we did was during the show was when have we talked about Lina Khan, the FTC commissioner? What did we say about her? And can you write a letter for her to appear at the All In Summit, Quote a couple of nice things you said about her on the pod. And it just did it right away. So, you know, producer didn't have to go looking, you know, spending hours looking up exactly what we said and you know, when.

Yeah. Wow, that's so powerful. That's incredible.

So this is a small example of what happens when you can incorporate all of your work context into the AI and into the chat and now it's just there and it can give you answers based on that. And this is clearly the future. And we've, we've heard this, I mean, ever since ChatGPT launched, what we hear over and over again from businesses and enterprises is how do we get that but for inside our company? Yeah, I mean it's not even that creative. Right. They just want like a chatgpt like interface for all of their work content. And this is why we think that, we think that interface belongs in the chat. Right. Because sometimes you'll start a conversation with AI and then want to add your coworkers and obviously you can't do that yet in ChatGPT or sometimes you'll be on a chat with your coworkers and you want to ask Blue AI a question. So you really want these things to be together, your human chat and your AI chat. You don't want to have to have two separate apps for this. Yeah, and that's kind of what we do is we bring these two things together.

Yeah, no, it's, it looks not only beautiful but. No, it's very well said in terms of just people want that and expect it. How do you guys think about like David, obviously with you know, with even Kraft being an investor and an investor in companies like ClickUp and you know, these workplace tools, even whether that's Slack or ClickUp, like the ability to have chat in different avenues of like the competitive landscape. You think about Slack, as you said it yourself David, at the beginning of, you know, people and companies obviously use the product. It's very sticky getting them to. It converts to something completely different. I know you, you know, you talked about the ability to import from Slack which we can get into. But I'm just curious how you view, you know, with this fast and you guys talk about it all in all the time, like fast paced AI movement. What do you think, you know, what do you guys focus on at Glue to really, you know, get in the market and do it in a way where it's like hey, other companies might do something similar but like this is our main focus point of how we're viewing it.

Well, I think we're a pure play chat app and you know, ClickUp is a portfolio company of ours and I'm on the board and it's a great product and everyone should use it and it. Yeah, basically. Yeah. And their value proposition is that's an all in one app that does everything from project management to you know, a whole bunch, you know, a wide variety of things. And I think if you want that all in one experience, you should definitely use that. I think we see what we're doing as more of a point solution that integrates and in fact we want to integrate with ClickUp, you know, very high in our list to integrate. We've never seen ClickUp or any app in that category as being competitor. We see it as a potential integration opportunity. Slack is a competitor. I think that is more like a head to head competitor. So is Microsoft Teams. Yeah, that's probably where we see the competition. Everyone else we Just want to integrate with.

Oh, that's. That's so great. And even too, I know I wanted to touch on a, the integrations that you guys are planning, but B's going back to the app itself, like the feed section and even some of the import aspects of what you guys will do, right? Like when as you guys are bringing on companies, talk to us about the importing and getting someone signed up if they have hundreds of employees, what does that transition look like or what have you guys experienced so far and what's been the learning in something like that?

Share my screen again and we can. Yeah, kind of keep going. So I'll just start here where we were at the thread and we have multiple groups on this thread and this kind of touches on what David was saying of how a thread can live in multiple groups at the same time instead of being stuck in a channel. Well, what you see here on the left, these are the threads I'm involved in. These are all the ones that I've started or mentioned in, pulled into, that I followed. You might have a lot of other conversations happening in your company. Maybe this is just a small percentage of what's happening, unless you're like me, CEO in all the groups. But that's really one of the core points of glue, is how can we really focus on the things that are for you and filter out all the noise. Well, sometimes you just do want to explore and see what else is going on. And so what we've done is embrace this concept of a feed, which is pretty common concept. It brings back harkens back to the yammer days of having a way to easily browse through conversations that may not be in your. In your inbox. So a lot of these are the ones that I haven't seen yet are here from our GitHub integration, actually. But if you see something that's interesting, you know, oh, I should have been aware of this. You can just follow it right there from the feed and then it'll show up in your inbox the next time you. You when you get an update from it.

So yeah, the distinction we see is that. Sorry, I didn't mean to talk over you, Evan. I think there's a little bit of lag there. Oh, go for it. So, so the, the distinction we see is that inbox is all the stuff you have to see that you have to respond to, whereas feed is really for discovery. Like when you have extra time and you just want to see what else is going on around you, you'll go to the Feed and kind of discover new things.

Yeah, that's great. It's. It's so funny. I mean, I being when Yammer came out, I guess I was 8 years old in 2008. Like, I never have seen this feed idea. And you know, when you say bringing back to the Yammer days, I think it's not only so powerful, but just like that high overview of what's happening in different aspects of the company and conversations when it comes to permissions, I think this is a question I just was curious about as I've seen you guys talk about the product, which is, you know, David or Evan, as CEO of, you know, Glue, when you're working with the AI, you know, what do, what do the permissions look like from, you know, an employee versus someone like yourself in terms of the capability to go back and pull data from the conversations?

So our permissions model is pretty simple right now. It's very similar, kind of to how email works, really. When you see that thread with the recipients, those are the people that can access that thread. So if you're in the group, if the group's on the thread, then those people can access it. And the same is true for, you know, myself versus, you know, an employee. Kind of have to be on the thread that lets you have private threads if you just want a thread with a few people. And then when you use Glue AI, Glue AI just takes your permission scope for those searches when you're in that private thread with Glue AI, when you add other people from the team and you start talking to Glue AI, what we do is we just pull information from the public scope that everyone can access so you don't accidentally share sensitive information from private conversations. So that's the way it works. Right now we're still iterating on that and making sure that people feel comfortable using blue AI and not exposing sensitive information. Of course, if you're an admin of your workspace and you want to do a data export or if you need to do auditing, of course we're going to have enterprise features that let leaders and admins do that. As a company, you own your conversations. But one thing that we really care about is that people can trust that the conversations that are private are actually private.

Yeah, no, that's. That's super powerful. And, and how about from the aspect of. I know, David, you talked about the example from the all in podcast and being able to, you know, feed the transcripts. When it comes to, you know, you're someone like yourself that has the fun, you're part of glue. You have so many portfolio companies. What does that look like in terms of like the overview of being able to gather data from different threads across different companies? Is that something that you guys see being in the future of what glue has to offer? Or will they forever be like siloed entities? Or can you kind of get like the, you know, the weak report across what's going on and key things that are happening or problems that are happening?

Well, I mean, I find it amazing because I'm in multiple workspaces. I'm in a workspace for Craft Ventures. I'm in the glue workspace, I'm in the all in workspace. So I'm in multiple workspaces and it's very, very good for. With Slack you've got that weird toggle up at the top to like the workspace switcher.

Yeah.

And there's no single inbox for everything. Whereas with Glue you have, you know, groups live in workspaces and you can be in multiple workspaces if you want, but there's just one inbox for everything. So there's only one thing for me to check. And whenever you have a. We also have flexible features around guest users. So you could add a guest to a group within a workspace or even to a thread if you want. And it will put like a yellow kind of alert lock thing to let you know that there's a guest there. So be careful what you share. But it's super flexible that way. And yeah, we've learned from both, you know, the yammer days and then also using Slack that people inevitably end up in multiple workspaces. So you want to have a more thought through experience around that as opposed to kind of like this weird federated switcher thing.

Yeah, no, I like that a lot. I guess just to change conversation here a little bit when it comes to the team itself, what does that look like today? You guys have been at this for three years, initially incubating it within Craft Ventures and now going out to the public talking about it bringing on users. How has the team grown so far and what does that look like in the future? Especially with what you guys talk about all the time. David is just with AI and with this being this tech product that you've had so much experience in building. And what does team look like now and what do you see that look like in the future?

Well, we think we can get a lot done with a relatively small team and I think we're taking a page out of a book of a company called Linear, which is also, I think achieved a lot with just a relatively small team. So we've kept it small till now and I mean, we're definitely hiring developers. That's just our biggest needs to keep growing the Eng team. But yeah, I mean, we think we can get a lot done here with a relatively small team and don't want to, don't want to get too big too soon in terms of staff.

Yeah, absolutely. You know, as someone that, again, like speaking from my past company, we had a team of 10, 12 people. You know, when you. Nothing massive, but when it comes to use cases for you guys, like, what do you, what are you feeling has been the best feedback? I don't even want to say the best, but just team sizes, they vary across the board. What has been the biggest gauge of interest thus far?

Well, one of the things that we realized early on is that if you have a product like this that is something that you use every day and that a lot of people rely on, you really need to have folks who are ready to give it a shot. And so we're kind of targeting early adopters. Honestly, we're targeting teams like us who maybe have been frustrated with Slack or other tools in the past and want to try something new before we grow too big. And so this even is reflected in the integrations we're building and the features we're building. We are targeting startups, we're targeting technical companies, people that are looking for the latest tools that are going to help them be competitive and it's paying off. We're seeing a lot of interest and it's clearly really resonating. We kind of want to be that tool. When you go and you look at the site and you look at the integrations page and you say, oh, this is, this is me. These are all the tools I'm using. So I can just plug it all in and it'll be better, perfect. That's kind of, that's kind of what we aim for.

Yeah. And I, you know, I just go back to the example of Craft Ventures. You know, we have about 40 people working at Craft now and so that glue works great for that. So it just gives you a sense of, you know, even for like a 40 person company, it works extremely well for us. I don't think Slack would work at all because we use Glue as our knowledge base. I mean, every deal that we look at gets its own thread and we put, you know, a bunch of information in the thread. It's really important to be able to find that thread later to be able to refer back to it and Slack. I mean, yeah, there's like threads, but they're really just in channels. They don't have names, you can't reference them. I'm not even sure you can link to them very easily. So like the, the content in Sock just feels much more ephemeral, which also is going to make it, I think, much tougher for their AI to work properly because everything's kind of munched together in these channels. Whereas, you know, for us, focused threads, that is how ChatGPT works. Right? Is you, if you change the subject, you want to change threads so that the AI doesn't get confused about what you're talking about. So in any event, for us, I mean, we need our chat tool to also be a highly organized knowledge base and that's what Glue has given us.

No, that's beyond powerful. A question that I, and we talked a little bit about at the beginning here. But you guys have been working on this for some years now and obviously having the background there. But why now in terms of officially launching it, when you guys did, you could have, I'm sure, waited an X amount of more months to redefine it more and continue making it better. And you talked about this concept of not just putting out an mvp, but why now in terms of the timeframe of taking in beta users and really bringing it to market?

Well, I mean, I think most people probably would have launched something worse sooner, to be honest. So I think we kind of waited as long as we felt like we could.

Yeah.

And you know, you don't want to wait too long because you're going to lose the market opportunity. There's really an opportunity right now that's been created by the need by every company, I think, is trying to figure out its AI strategy and how to bring AI inside their company. And again, our hypothesis is that AI should live company. The company's internal AI should live inside the chat. And so we want to get to market with that, you know, with, with that product. And, you know, we felt like we waited as long as we could in terms of making it feature rich and really polished.

Yeah, absolutely. Yeah.

I don't know if you have anything.

Else to add to that.

Yeah, look, I'll add to that. I think frustration with work communication tools has been building for a long time. You've heard of channel fatigue or slack fatigue for years, really, but I just think it's been growing to the point where people are actively looking for something new. I'm seeing tweets all the time, like, hey, is there an alternative to Slack? And I've tried Discord, but it's not working. It's not made for work, and on and on. And so I think it was this combination of us seeing the opportunity with AI along with, hey, we've kind of just reached this breaking point really with the alternatives that are out there, that it's a good point for us to jump in and say, hey, here's something better. Here's, here's a. Here's where we think it can go.

Yeah, yeah. And I think the bigger your company gets, the more you experience those growing pains with Slack because again, you have this problem of over time, on a long enough time horizon, every employee ends up in every channel because again, you just want to get to that one threat. So think about how that multiplies. The complexity becomes exponential. So as you get into hundreds of employees and thousands of channels and everyone feels the need to be in all of them, it just becomes very unwieldy. So we think this is going to appeal to companies that have hundreds of employees and eventually thousands of employees. I think you start with the early adopters. And so, you know, our most common company right now is, you know, 10 employees, 40 employees that, that kind of size. But we think we're getting to the hundreds of employees pretty quickly because again, they're the ones with the most channel fatigue and most frustration with Slack.

Yeah, and that's awesome. David, with your background of not only being in tech, but having a fund, investing in incredible. So many great startups. When you think about the future of glue, and this is more so to both of you, but David, first, what excites you the most about what you're building here? And there's so many directions and things that with the emergence of AI and how it's changing so fast, what excites you the most about the product today and what it can be and what it will become in the future?

So first of all, just getting to work on products for me is really fun. And I'm not just looking at investments all day and writing checks. It is actually fun to geek out on products. And Evan and I do go really deep, you know, discuss every little nuance of this product. And sometimes we disagree. And then finally when we actually reach a convergence, it's actually great. Like, we had a meeting like that today where we got on the same page about some things we've been, you know, haggling over for a while. And it's just, it's really fun when that happens. So, first of all, it's just fun for me to work on products.

Yeah.

Beyond that, I think that we have a really great opportunity, like I said, to kind of be the home for AI in the enterprise, to be the main place you go to interact, to kind of be that knowledge base. And again, it's because that chat history is such a rich place to get information, to get the most current information. So, for example, like Evan showed when you do a prompt on who's the expert in Fill in the blank. You know, I've done that at CRAFT for, you know, who knows about this company or who. Who's our expert in, you know, marketplaces? I can ask that question. The AI will give me a really good answer. Finding the expert in a company is a. Is a really big deal. And in fact, there's lots of enterprises spending millions of dollars every year trying to build these corporate knowledge bases, I think.

Yeah.

You know, and employees have to, you know, fill out these elaborate profiles, putting all their skills and expertise on there. All that just goes away because the AI is going to figure it out based on your actual contributions to the chat.

On that note, too, and this is maybe something that I'm just thinking about. If you have a team of 10 people and you say that you can upload transcripts from all in, could you upload bios about yourself, your skill sets and things about each employee that would then gather that rich knowledge to then be able to pull back from. Or is it more just based off of the communication that's happening over time?

There's definitely some of that you can do now. You can add group descriptions, you can add bios and information to your profile, and that's all available to Glue AI as well. There's a lot more that we're thinking of, what we want to do there. But really this is all about having what is the latest information that's being talked about out there. I'm sure you've been in companies where the same question gets asked dozens of times by all the new people joining, or maybe there's only one or two people with this little arcane bit of knowledge about this system and it's not documented anywhere. Well, suddenly you can actually take advantage all of that, and maybe even automatically, maybe Glue AI could even sort of realize that that's what you're looking for and suggest it to you.

Absolutely. Well, before we wrap up here, guys.

By the way, just to add to that a little bit, I think this is one of the areas that will make. Let's Call it consumer AI, different than business or enterprise AI, is that what we found is there's a lot of value when you give the AI context to know, like, okay, these are the, these are the people in my company. This is what my company does. You know, these are groups. So just giving the AI the awareness of objects in the system, like will dramatically improves the performance.

Right.

If you just went to ChatGPT and asked it a question, it's not as good as if it actually understands, okay, this is your company. The AI knows that I am this company. We have these employees and so on down the line. And so we're going to keep building more and more stuff there to help structure that problem for the AI. The AI is really good at reasoning if it knows exactly what to look at and if it has, again, the right context. So making sure it always has the right context to give you the best answers is something that we feel like we can go really deep on. I think there are a lot of people who are trying to solve this problem in different ways, but again, we think that the most natural place for this to live is in the chat.

Yeah, I think that's the big takeaway. Context, context, context. And even for me, I, I always, I read into all these, like, personality tests, like, I'm sure, like, if you could pull in data from personality tests across your team to like, really figure out, as you said. Like, I saw this thread one time about how in this company it was, you know, they gave, right, an ad to all the employees and the girl that was doing the social media management ended up writing like the most compelling ad that converted the best. And it's like being able to find these things that you might not see using, you know, things like more context is extremely powerful. So it's great to hear and I'm sure a lot of question that all of the entrepreneurs and founders listening are thinking, like, how do I get access to Glue AI? What does that look like right now in terms of onboarding companies getting signed up? I know there's a big wait list here, so let's dive into that before we wrap up.

Yeah, sure. So there is a wait list. We're churning through the wait list quickly. We've got a team of people giving demos and onboarding calls. You can add your name to that on a wait list. Just go to Glue AI and we've got a very simple email form right there. One click, you're on the waitlist. We're going to be accelerating through that as quickly as we can. And like, David mentioned. We're hoping to just kind of open the floodgates as soon as we can. So. Yeah, do it. And of course, we'll keep you updated on your place in the line and updates on the product as well, if you put your info in there.

Love it. Well, Evan, David, I just want to say thank you so much for coming on the show today. I'll be sure to link everything down below. And I got to say, I'm excited myself to get myself, my team signed up. So pumped for that. Pumped to start using the product here in the short term. And thanks again for coming on the show today, guys. Really appreciate it. Yeah.

Thanks for having us.

Yeah, thanks.

Thanks, Casey.