Latest / The Chief Customer Officer Podcast / EP009: Sitting Ducks and Fortresses: How to Read the AI Landscape

Transcript

EP009: Sitting Ducks and Fortresses: How to Read the AI Landscape · The Chief Customer Officer Podcast

• Jeff Breunsbach: All right, welcome back to the episode of cheap customer officer.io. This will be coming out on Tuesday, March 12th. We're almost Thursday, March 12th. are currently, yeah, Thursday, March 12th. I was going to say it's, we're almost a third of the way through the year. You know, I don't know if that scares you or if that is, you know, if that's motivating, how does that land on you? Jay Nathan: Thursday. Yeah I was talking with somebody about that just the other day and our fiscal year is February 1st. So by the time you get two months into your fiscal year, you're already like a quarter of the way through the actual calendar year. It's like, man, it's by it's crazy, but you know, for the young young people listening, it's all relativity, right? Like the older you get, the faster they go. ⁓ Jeff Breunsbach: Alright, what's the? right, give me the TLDR and why do you do your fiscal year in February? Is there a big benefit? Jay Nathan: do that to match our primary partner, Pendo. Because we can match their intensity at the end of the quarter, it just makes easier. So we report by that. And ⁓ there some practical reasons for doing it too. Like number one is you're not doing financial close the year and closing your sales year right in the middle of holidays. So I mean, I've worked for companies that have been on a calendar year, it's sort of brutal. Jeff Breunsbach: Cut it. Okay. Yeah, makes sense. Yeah. Jay Nathan: ⁓ and so people get to enjoy their holidays a company that closes books and it's, and sales year in January versus December. Jeff Breunsbach: This is very true. I remember a handful of times negotiating deals between like Christmas and New Year's and you're like, this is like, know, no one's working now. And then you're trying to call people like, and then you feel bad because like, I got to call you while you're on your vacation. But I need you to sign this doc, you know, it's our CFO or whoever our CRO is, you know, out doing something. And you're just like, all right, I got it. need you to, know, you're on, you're on Can you look at your iPhone for the doc you sign and click the button? You know, you're like, Jay Nathan: Yeah. Yeah, right. Yeah, who's going to do that? Right. And it's, I mean, it's riskier for them, right. Because they get all the sign-offs they need on their end. Like, so just makes everything better when you don't have to do that. So I mean, you still have the end of month you got to deal with. That's still a thing, but it's definitely better. So. Jeff Breunsbach: This is crazy. Yeah. Alright, so we got it. think we each have a couple articles that there are a couple LinkedIn post articles that kind of got us thinking that we wanted to bring to the table and also sounds like you have built something and wanted to showcase it. Where do you want to start? Which one you want to go to first? Jay Nathan: Well, actually, start with one of your articles because I'm trying to find where I write this that I have with Claude where I built this thing that I wanted to show you. And it relate to one of the articles that I wrote or that I picked up. But ⁓ actually, I to tell you something really interesting first. And I think, ⁓ you know, lot of people, a lot of people looking for jobs right now, people are feeling some displacement, lot of layoffs seem to be happening still. Jeff Breunsbach: Sure. Jay Nathan: And I heard somebody ask the question their day, like, do I need to do? Like, how do I get into an AI first kind of company? How do I show that I'm AI first? And I think the answer is you got to be AI first. had that I've talking to about coming to work at Balboa for a few weeks now. She's in the interview process. Yesterday, she ⁓ sent me an on what she's been working on, and she actually sent me an app that she built in LUMmable. She's not a technical person. got a marketing background, has some Pindo experience, pretty well-rounded, very interesting person. ⁓ she sent me an app that she built and she's like, hey, check this out. Here's what it does. Here's how it relates to what you guys are trying to do at Pindo. And is what you got to do if you want to get a job. And I would say that's like just scratching the surface, right? But prove that you are getting your hands dirty with AI if you want. Jeff Breunsbach: Yeah. Yeah. Jay Nathan: to work for an AI first company. And I mean, we consider ourselves AI native because we're thinking about it and building every day first, but like, that's how you do it. I was just really impressed with that, you know, so have you seen people doing that? Or are you thinking like, Jeff Breunsbach: Yeah, yeah, well, like, ⁓ you said that there's like three things that came to my mind. So the first is, I think it's like a David Goggins quote or something, you know, motivational speaker that I've seen recently, which is basically like, you have to think, you basically have to that you are the person you think you are, even if you're not there yet. And so like, like you said, like, if I'm not doing AI stuff right now, have to, I have to get to my desk and at least you know, start to think, okay, like, how would I do this AI first? Like, what am I going to do? But like, start putting yourself basically in future Jeff, which is like, I'm already AI first, like, you know, so like, I think like, it's a mindset or like, you know, I feel like there is some motivational pieces that, know, you talk about kind of working out and eating healthier and everyone's, you know, New Year's resolutions always fall by the wayside. And it's like, well, it's because it's you can't just do it for a period of time, right? It is just the person you become. So like, anyways, that's, that's like the first thing that came to my mind. The second is Jay Nathan: Yeah, I am. I'm that guy. Right. Jeff Breunsbach: love that she did that ⁓ kind of sent you something that she built. think that's totally a way to stand out. think, I think also like, I mean, I think like that's almost like to the extreme, right? Where you like, she went down and built an entire app around this. But I think like, even if somebody was, where it was able to send me a loom video, you know, just a Google doc that they wrote up, but like something that would articulate that like, Hey, Here's how I would approach it by using AI or here's like a workflow that I'm thinking about or anything, right? But I think like even like, I think like building the app is like the extreme example, but like, I think there are steps that you can do along that route that like are still, you know, showing that you can do AI skills. And I think like that's a big part too. And then the third, that I think about often ⁓ I actually think this came from, it might've come from you that you are the, you are the CEO of your own career and Jay Nathan: Yeah. Jeff Breunsbach: The thing that has always resonated with me is that like, do anything. I mean, like it is your career, right? So like you can do anything you want. So there's nothing stopping you from, like you said, like, well, I just filled out the application on the website. Cool. Well, there's nothing stopping you from finding who the hiring managers are and sending them videos. There's nothing that's stopping you from building an AI tool that you think they might find useful. There's nothing that's stopping you from. recording a loom video about like why you're the best candidate and sending it via email, right? Like there's just like, if you play by the rules that are, if you play by the direct rules, like you're just filling out applications along with like 30,000 other bots. But if you start to think about, okay, I'm the CEO of my own career. And if this is like the job that I want, and this is like the deal that I want, then like, how would I go get it? And there's things that you can do outside the rules that, you know, I don't know, aren't illegal. I don't know that don't like Jay Nathan: Yeah. Jeff Breunsbach: puts you in a negative light, right? So I just feel like there's that, if you kind of come with that mindset. I don't, those don't relate back to AI. They're more like, I don't know, philosophical life things. But when you started talking about it, like that's just kind of where my mind landed. I think like that AI example kind of fits into that, like a couple of those buckets. Jay Nathan: of all, I don't know how you came up with three such organized thoughts that quickly. So ⁓ going to chalk it up to your 10 years younger than me and your brain is much more functional than mine is at this point. ⁓ impressive. Very impressive. Atomic is one of my favorite books. What's guy's Jeff Breunsbach: Coffee and creatine. James Clear? Jay Nathan: James Clear. Yeah. And, he, he talks about that a lot too. It's like, just if you, if you just tell yourself, you are the person that, you know, you're, you are a fitness guru that gets up and works out every morning. You are AI first. Like just, you got to start with the belief right to your point. ⁓ but that last point you made is so, so valuable. And I think so many people just sort of play, try to play by the rules and sort Jeff Breunsbach: Yeah. Yeah. Jay Nathan: you know, follow the path. Yeah, I apply for a job. I'm not getting any callbacks. Well, no, no kidding. You're not getting any callbacks because everybody's applying for jobs, right? There are thousands of people like to your point, you got to really get in there, roll up your sleeves, go network, break the rules, do something different. So you get noticed. That's like a good, like a good business lesson to like a good marketing and sales lesson. Like don't just do the thing that everybody else is doing. What people don't realize is that your career. You've got to market yourself, especially these days where there's just so much noise out there being generated by other candidates and people that are going above and beyond, people that are trying to stand out. And you're not going to stand out by just putting your resume in a pile. Anyway, this isn't about, you know, recruiting or finding jobs, but those things definitely, definitely resonate. Yeah. So. Jeff Breunsbach: Yeah. Stand out. Uh, well along this AI route, I think one interesting article that I've found is, um, by Kyle Norton view. Do you know Kyle? Have you heard or seen him heard of him? Um, he's the CRO owner.com owner.com is a restaurant platform. Um, that's become relatively successful. It sounds like, know, from the outside, from the articles that are written. Um, but he had a, he had a, um, LinkedIn post. And the thing I thought that was, was great was just his first. Jay Nathan: ⁓ yeah, yeah, yeah, I have heard of him a good bit actually. Jeff Breunsbach: question, which is basically every revenue org is making one of the same decisions right now, which is, do you let AI adoption happen organically across the team, or do you start to centralize it into a small group that builds out for the entire org? And I just, I find this interesting because like I experienced this actually at, at spring health when I was there, you know, we, were a 1500, 2000 person organization. I mean, you start thinking about, right? Like everyone was using AI and all sorts of use cases in different ways and in You want that to happen, but then you also pretty quickly realize like how. It all has different contexts. It's all using different things. We're all saying different things now. Like it actually has like almost like a in some cases it might even have an adverse effect where it's like cool. We're adopting tools, but it's actually making us either less efficient because of the outcomes of those tools or it's making us ⁓ less on message on brands. Like there's just so many, you know, kind of directions that it could go. ⁓ And now we're you know, Matt ⁓ much smaller. company in Junction. And I think we're starting to think through this right now, trying to figure out, know, we all kind of use tools every day and engineering is using stuff and CS is using stuff sales team. And so I think we're starting to try and ask ourselves the question of like, do you start to centralize it now? Like, is it best to centralize it now while we're all still nascent and early and like we can, you know, kind of push this into a central. So ⁓ I don't know, doesn't, maybe there isn't one answer, but I thought it was just an interesting question. And I'm curious if you've seen this with people you've been talking to recently and, ⁓ in some of the organizations that you all have been trying to get into, like, you know, do you find companies doing it one way or the other, or do you think it's gotta be some mix of both? Jay Nathan: I was looking for some notes from, forget who I was listening to the other day while I working out, it talked about, ⁓ it was very relevant to this topic. And I think the funny thing is there are still so many companies out there that haven't even approved these tools for just day-to-day use within the organization by the individuals. They're probably doing something. you know, centrally. But my guess is that they have, if they haven't approved these tools at the enterprise level, they're probably doing the organizational stuff very slowly too, right? Either the product development or like the internal tooling. So I think you need both right now, especially because what we're trying to do, and actually we have another, it's time that we have another AI show and tell coming up. That's our internal like, Hey, here, let's, I'm going to show you mine. You show me yours. And, ⁓ Jeff Breunsbach: Yeah. Yeah. Jay Nathan: The reason we're doing that is to first keep the fact that everybody needs to be experimenting and pushing the edge here a little bit with AI, top of mind for everybody on our team, every single individual. It's key for us because we're advising our clients on AI solutions in a lot of ways. ⁓ So I think yes to both. ⁓ Jeff Breunsbach: Yeah. Jay Nathan: You have to let AI adoption happen naturally across every person in your team because a lot of this is departmental. A lot of the efficiencies that you're gonna gain are gonna be departmental in nature, very specific to the tasks that your team has to execute on any given basis. But the centralization is also key. And more and more, I'm seeing that if you don't have everything centralized, It goes back to our conversation last week, Jeff, about the poor experiences that we both had two weeks ago with support, right? The reason those support experiences sucked for us, the agentic support experiences, just to catch everybody up, Jeff and I both had independent two different companies. We were trying to get support. We both got channeled to their support agents, like agentic AI agents, not people. And, you know, both ended up, you know, but frustrating us and then eventually shelling out through human anyway. So it's like the worst of both worlds essentially. But the reason those experiences were poor in my opinion is because at least mine, that experience, it was not informed by the whole context of who I was as a customer, what the product was doing and sort of the challenges surrounding that particular issue that I had. Jeff Breunsbach: Yeah, yes. Jay Nathan: And so it didn't have enough context to do the right job. The tools are smart enough to do it, but if they don't have the right context, they can't do the job. that takes, I guess my point there is that that takes centralization. So you have to do that too. You have to centralize the data layer for these applications at the enterprise level. Jeff Breunsbach: Yeah, I think where I was, I think like I'm, I agree like that, like both are needed. And like, I guess the, like the way that I, I didn't really draw this out or anything, but like the way I kept thinking about this is like, it feels like pretty quickly, there are tools that you're allowing access to context of your systems. Like it's plugging into MCPs and HubSpots and whatever. Like to me that that's the stuff that seems like it should be centralized because you need controls. You need like, ⁓ you need. ⁓ user credentials and other things that like you should probably centralize those types of things, but then you still should allow your teams to go, ⁓ you know, mess around with tools that are in their department that maybe not are, you know, aren't touching main systems record. It kind of feels like to me, like you're starting to centralize around the, like the check GPTs, the, the clods, like those feel like, okay, we're starting to centralize on those. Like those should have context. They should be connected to our tools, but like there needs to be like tight level of controls and permissions. And then like an example I was just thinking about was like gamma, which I don't know if you've heard of gamma, but it's like a, ⁓ you know, it's a tool that allows you, yeah, presentation, it'll create docs and other things, but you can essentially feed it text and then it just builds you a beautiful looking document is the best way I can say it. ⁓ And so to me, right, like that's not touching, it's not touching our main systems of record, but like I'm using it from a CS perspective and like, cool, I should keep playing with that because that's getting me efficiency in my department. It's a $20 tool a month, right? Like there's. Jay Nathan: presentation tool. Jeff Breunsbach: I don't know. So it feels like you said, it feels like they're starting to become buckets of both. And like, you can start to see how some of these tools that are touching main systems of record, like they're, you know, there needs to be some guidelines and regulations around it so that you don't, you know, all of a sudden delete all of your records from HubSpot, right? Like that would be catastrophic or terrible. You'd have to like, you know, that'd be a business defining thing. And so like, it feels like, I don't know, you're starting to come to both, but I agree that like both need to happen. And I thought it was an interesting question that he posed though of like, ⁓ of this. The other, kind of goes on further. We don't have to talk about this part. I didn't find it as interesting, but he basically was like, you know, ⁓ there's somebody he was talking to that basically is like, ⁓ a job is just a bundle of tasks. And so then like, you can basically pull apart any JD and ask it like, what's done best by humans, what's done best by jobs, which I like start to generally, I think like, that's a direction that we're all going, which is basically like, okay, where, how can I get my team doing the best work? Jay Nathan: Yeah. Jeff Breunsbach: And like, where is that versus like, what can we repeat and automate? I think that's like a pretty standard question that people are asking these days. Okay, the. Jay Nathan: I saw another LinkedIn post the other day talking about, I think it was a customer success role. it's like, look at this role is going away. And it's like, look at all these different things that the agents are now doing. it's like, well, you know, those are all tasks. Like don't confuse a job with a bundle of tasks. Right. And yes, these tools are taking tasks off people's plate. And yes, like we talked about last week, AI and intelligence is being used as air cover to basically conduct layoffs right now because people are companies are still bloated from the pandemic. There's no question about it, but don't confuse that with like a replacing a full human being. That is a lot different and I don't think we're there yet. Jeff Breunsbach: I think we're pretty far off. All right. The second one, ⁓ Kyle leads a sales team at LaunchDarkly and he also has his own little, I think his website, I call it little, did not mean little, his own website called Sales Introvert. he runs, he puts out some great content about ⁓ kind of being a sales leader, targets, trying to assign territories, like how do you get the best out of your sales reps, that kind of stuff. ⁓ but I thought this was like interesting, like his first kind of, he kind of listed three, ⁓ decisions with AI and go to market. And like, I thought these were pretty on par, ⁓ with like, I think we've kind of alluded to this, right? Like everyone just thinks, okay, I just need to run and start automating everything and AI and everything needs to have AI without really starting to think like route without really stopping to think. And I think this is the part of his post, which is basically like. Number one is automating bad processes at scale. Like you actually don't know if your process is good right now, the way it is. ⁓ And I've actually started to ask Claude and ChatGPT, here's the process I have today. I'll either draw it out or I'll just list out the steps. And then I'll say, you know, like here, let's, let's actually go through and optimize this. what, you know, where, where do we have efficiencies? What can we do? How do we change this ⁓ to try and figure out like, where do I have opportunity there? So I think that's an interesting one he listed. ⁓ Building on generic best practices, like ⁓ you know, he kind of lists in here, you know, people are trying to become these prompt engineers within these tools and saying, ⁓ you're, know, you're the top sales person. You're an elite seller. Like, ⁓ but it still is, it still is then trying to go figure out what is like, it's trying to go learn what is an elite seller, but that's based on context that it's getting from all the data it's trained on. It's not, it's not, what does it mean to be an elite seller at launch darkly? Or what does it mean to be an elite seller at, you know, this B2B SaaS company and this specific vertical. And so like, Jay Nathan: Yeah. Jeff Breunsbach: Again, you're going to those tools, again, think about, I guess, the context that it's trained on, which is these generic best practices, which I thought was interesting. ⁓ And then this one I thought was just one that would probably warm your heart, which is prioritizing your efficiency over the buyer experience. And I see this right now, I actually go back to our support experiences, right? They're optimizing efficiency right now for us, rather than trying to think about what our experience is from a CX lens. And I think, again, Jay Nathan: Yeah. Yeah. Jeff Breunsbach: It seems like we're jumping so fast into, me just turn this into AI. Let me turn this into an automated workflow without actually thinking like, okay, are there like rip cord moments when if it's going wrong, we can just pull the rip cord, put a human in the loop and like, almost like stop that before it gets even worse. Right. And it just feels like right now those don't exist often enough. And we're just kind of letting the system run. And then at the end, looking at some data saying, ⁓ I guess it didn't do it. Let's train it to do better versus. Jay Nathan: Yeah. Jeff Breunsbach: Hey, you know what? It got it wrong the first time. Let's just put somebody in before it gets more frustrating. And then why did it get it wrong the first time? I don't know. It just feels like there's some experience things there. Jay Nathan: feel like you could take all three of these points, go rewind the clock four years to when you and I were doing gang row retain. And we would have said the same exact things, right? Automating bad processes at scale, building on generic best practices and prioritizing efficiency over the buyer experience, which didn't we talk about that last week? The service blueprint versus yeah, we actually published a newsletter on it this weekend. ⁓ but Jeff Breunsbach: Yeah. It's true. We did, yeah. We wrote about it. Yeah. Jay Nathan: I think like the generic best practices is so funny when you scroll through LinkedIn. These days you see like these people posting these infographics. You can tell they've come over from Twitter and they're like, here's, here's the, your complete guide to X, Y, and Z. And it's like, well, I can tell that you just generated this with, you know, whatever it's not actually based on your experience. First of all, cause it's some, sorry, but it's like some 22 year old space on the, on the post. Like you don't have all that experience, right? It's fine. But Jeff Breunsbach: Yeah. Jay Nathan: Like it's just overload. It's AI slop. So I completely agree with you on that. And by the way, like you don't need to use generic best practices anymore because like if you don't have every call recorded by now you're way behind. So I'll just start there. Like that's, that's price of entry at this point. Think about all those calls that you have with customers, with prospects, like you've got your own corpus of ⁓ raw data that can now be interpreted. You should be able to really quickly figure out what the best practices are for selling for retaining customers. Like what are your customers core problems? Just, just off that one data source alone. You don't need anything else. And by the way, if you're using a tool like we use for, for call recordings, you can go do an AI query across all of it today. Not efficient, not fast, but you can do that. So, ⁓ yeah, this is pretty cool. I just don't think, I don't think it's really new. I mean, I think it's like, you know, it's. Jeff Breunsbach: It's yeah, I think it's like you said, I didn't think about it this way. But I think like to your point earlier, it feels like these are manifestations of what of things that already existed. But now you actually have to put this on top of AI as well. Right? Like it's like, we actually weren't doing these things before AI existed. Like we were automating bad processes. were, you know, using best using generic best practices, like, but now it's actually even probably more exacerbated because you've just got like more slop, more ideas, more stuff coming from AI. And then you're just thinking, okay, let me layer this all on. ⁓ So that's a good point. Alright, those were my two articles or two LinkedIn posts that kind of got me thinking. think I saw those on Friday and over weekend. Jay Nathan: had one more thought and it just left my brain. So I'll just move on. So all right. So the thing I was going to share, we've talked a little bit about AI disruption in SAS. I don't know if you read, what's his name? Tomas Tungas. He had a really interesting article. Jeff Breunsbach: Thomas. Jay Nathan: By the way, I'll say this, my bet on IGV, it's only been a couple of weeks. IGV is like a software index fund. It's starting to pay off already. I do think it was oversold. So I'm gonna keep everybody updated on that. I actually bought more of it yesterday. you know, again, not giving anybody else advice here on how to invest their money, but I'm really curious to see how that plays out in the long run. But this is interesting because what he was talking about, you know, I think it's like, Is AI gonna kill SaaS is the question, right? We've seen a lot about that. We've heard a lot about that over the past few weeks, especially, you know, with the sort of this little crash in the markets. But the point that he's making here is that the answer to that question is actually very, very nuanced. Okay. As you would imagine, right? It's not, there's not one like AI killing all SaaS. Well, that doesn't make sense, right? I mean, you've got companies at different levels of maturity. Jeff Breunsbach: Thanks Yes. Jay Nathan: ⁓ different categories of software that ⁓ have different susceptibility to disruption. So ⁓ he sort of made this call and I actually put together, now I'm going to share something different with you. Let's see if I can do it. I created a two by two based on his ideas here. I'll show you my clots. Sorry, I'm like all up in the camera there. ⁓ ⁓ Can you see my clod now? Jeff Breunsbach: ⁓ yes. Jay Nathan: Okay, so this is something that I created called the AI vulnerability matrix is basically a two by two based on on this article. So the idea is like which which products are most susceptible. And the reason I want to talk about this is because I want people that are listening to this us talk about this. Think about what kind of company you're in. It might inform either what you feel what you need to do with your company to be successful or it might inform what you need to do as a professional to get onto the next thing. Okay, so that's really the why behind this. But the way to think about this is complexity of the solution itself, right? How hard is that technology to replicate? ⁓ sorry, how complex is the solution for the organization to own and then how hard is it to replicate? So those are the two axes of the two by two. So for example, if you're low complexity and easy to replicate, then you fall into the sitting ducks category, right? And probably that's always been the case, but even more so now with AI. Categories might include like basic sass tools. I think about like screen recording, right? Loom, for example, which Loom smartly is now part of Atlassian is part of a bigger platform. So the play there, yeah, it's part of a suite. The play there is to figure out how to go get wide, get bigger, you know? Jeff Breunsbach: Sweet, the products, yeah. Jay Nathan: cover more ground as a solution. ⁓ But you may want to consider is that the kind of company you want to be in right now, right? If you're a point solution ⁓ in this kind of world, everything trends towards platforms. And I think it's even more so now. So then you've got low complexity. So the solution is low complexity and hard to replicate. That might mean you have a data asset or some other kind of moat around. that product. like infrastructure, ⁓ again, you have some kind of data source. It might be easy to replicate the tool itself. Think about Apollo. I could go create a call sequencing tool or SalesLoft, like an email sequencing tool. I could do that immediately. What I can't do is go collect all the data about people. Apollo and Zoom Info have huge infrastructure and teams built around going to collect lead data and put that in a structure and a format so that I can access it as part of their services. And those services are still highly valuable, right? Because I still need to contact people. Jeff Breunsbach: Yeah, I think another category that comes to mind for this too might be something like if you're selling into like the government, because it could be an easy solution to replicate, or it could be a low complex, low complexity, but it's hard to replicate because governments are, you know, typically slow buyers or hard buyers to get into you have, you know, rules, regulations, they've got to do budgets and certain like planning in certain ways in terms of like how you can get like something across the line. So it could be another category that like, Jay Nathan: Yeah. Great point, great point. Yes, that's a fantastic point. Okay, cool, and by the way, I also had it give me departmental level impacts of what you need to be doing. So maybe we can publish this on the website. I think this would be pretty cool. All right, so then you have high complexity tools that are really complex for the customer to adopt, but they're easy to replicate, right? Jeff Breunsbach: That's cool. Yeah. Jay Nathan: For example, that might be like enterprise CRM or financial platforms. mean, they're complex systems, you probably could replicate those pretty easily. So these are targets for, see a lot of AI company, like AI native companies springing up to grab market share here in the accounting space, ⁓ in the CRM space, know, think ERP for the most part, complex software, but. everything's easy to replicate at this point to some degree if it's just feature function. And then, and then there's the fortresses, right? The ones who have both like high complexity solutions for the customer, but very hard to replicate. So maybe it comes back to those data assets and other, other modes. So anyway, what do you do? Like, I think, you know, you'd love to get into a fortress builder if you could from an employment. perspective, right? But either way, you can be successful in all these. think the strategy is just, is just different depending on which quadrant you find yourself in. And half of the battle is just knowing where you are on the field, right? So you can play, play the game you need to play. If you're an executive at one of these companies, it's like, well, we might be a sitting duck, but here's the strategy for going to win. If we're a sitting duck, like we're going to go. Jeff Breunsbach: Yeah, that's what I was gonna say. Jay Nathan: build partnerships, we're gonna go innovate as fast as we possibly can, ship every week, right? New capabilities, make sure our customers know about those. ⁓ And then, you know, basically buy ourselves more time to go, ⁓ to be a going concern. So what do you think of my, by the way, I didn't even have this matrix generated yet and I did this while we were on the call. It created this visual thing. Isn't that cool? Jeff Breunsbach: Yeah. ⁓ nice. Yeah. Yeah. Yeah. These two I mean, this is like the little I put together that, you know, tools of C era AI and CX like, and it did the same thing, you know, created some HTML that allows you to just it. I mean, I think it is like spot on I read that article. And I like, I think coming up with a two by two makes a ton of sense. I thought, you know, he also in that article made a couple interesting points. You know, I, about the, I think about like his perspective and bringing some of the data behind it, you know, like I think looking at the chart about GitHub and actually how GitHub had it like this, it looks like outsized advantage, right? Like they had 20 million users, think using, their product and kind of like overnight, like Claude and, and open AI turned on their tools. And it was like, ⁓ and you know, it was almost like this inverse. You could see how, like how much it goes down. And so ⁓ I thought it was pretty interesting that like, you know, he was using that as an example to basically say like, that was Microsoft, who's one of the largest companies in the world who probably has a huge balance sheet, you know, and like they couldn't even stop this from happening. Like sometimes this is, you know, this is how ⁓ the public, you know, adopts tools. And like, sometimes you just can't get behind the groundswell that this is creating. Jay Nathan: Yeah. Yeah, that's right. The sword of Damocles. I had actually looked that up. It has to do with like a sword hanging over somebody by a horse's horse's hair. And I guess it's a medieval analogy there. But but yeah, I mean, to your point, like GitHub co-pilot, that was the number one thing like code completion, like auto completing a line of code was the first use case for AI, agentic coding. Now it's like, why would you even look at the lines of code, right? You're generating Jeff Breunsbach: Yeah. Yeah. Jay Nathan: thousands of lines of code without ever seeing any of them with Claude code and these other tools. And by the way, it's just part of this is the disruption cycle is just so much faster right now because the innovations are happening so quickly, ⁓ which is fun, but also terrifying at the same time. Right. So, ⁓ interesting. Okay. Cool. So that was, that was one. Jeff Breunsbach: Yeah. Yes. Jay Nathan: The other one, I'm not even gonna share this article on the screen because frankly, ⁓ I tell it was AI generated and it didn't sit right with me how much AI generated it was, but there are three points in here which I are interesting. So there's a lot conversation right now about like, what does an AI first company look like? And so this article was sort of commenting on Jeff Breunsbach: Sure. Jay Nathan: Jason Lemkin post last week about what an it wouldn't agent when an AI first company is There are three things that they cited in here that I think are pretty interesting number one is just the sense of urgency versus ⁓ Do you have anxiety or urgency around? AI and I think if you're AI first, so I'll the three out then you comment and tell me if you agree with these You you see urgency around it. You're not fearful of it, right? Duh, that sort of makes sense. Number two, and this is an interesting one maybe for our audience, is the shift from CSM to forward deployed engineer. We can dig into that in a minute, but like of the points that was made in here is really good. We've been building customer success teams to manage relationships. True. native companies are deploying forward deployed engineers to drive adoption and solve real problems. One model assumes the product needs a human buffer. i.e. customer success, the other assumes the product should speak for itself very quickly. I couldn't agree with that more, right? Like I think the age of relationship-based retention and selling not but it's going to be challenged, right? And you're going to see a lot more value realization, actual realization. as part of the equation. that's number two. And then the third is the 700K AAR benchmark per employee. for context on that, four or five years ago when we would look at a prospect, Jeff, when you and I were our last company, we'd look at prospects. didn't privately held companies. ⁓ to estimate their annual revenue. would say, okay, how many employees do they have? Which you can easily find on LinkedIn or Zoom info. Okay, they have 500 employees, multiply that by $200,000 and get roughly their size. So they'd be a hundred million dollar company. Okay. That apply anymore because if that's your benchmark 200, then you're probably growing slowly. It's a big because your competitors are probably nearing this 700, 800, 900, 1 million per full-time employee benchmark. So I actually think that's right because you're delivering more to customers with fewer humans in the mix. And so I think that that metric is really spot on for AI-first companies. And in fact, I think there are some AI-first companies which blow that number out of the water. Jeff Breunsbach: Yeah. yeah, I mean, think largely agree with one and three. think I'll, just for the purposes of a podcast to play the devil's advocate for number two for you about relationship selling and kind of relationship. I think humans are going to be in the loop of buying software for much longer than we think because we already how bad the tools are today, right? Like I was reading an article the other day that I thought was pretty interesting was like, there's a lot of stuff that's being put out about, I built this website, I did this, right? I can do these cool things with it, but no one's talking about like it, like completing a whole, like you said, whole person's job, an end to end function, being able to really have enough context in order to like literally replace an entire human. And so I just think that that That future to me, know, now it's like ⁓ whole, I think the whole game is around, well, agents are going to be buying from agents and humans are going to be out of the loop. And again, I could see that, but also like these solutions don't work today. Not saying that they won't work tomorrow, but like, I think they are not working today. And I think in perpetuity, the way that people are presenting or thinking, right? Like, I think there's a lot of marketing going on around these things and like what their capabilities are. and what it means, You're a lot of these companies are forward selling what they can actually do. And I think like the future of that, I think is actually further than we think. And so ⁓ to point, I think going to happen. But I think like that shift is going to actually think is that shift is going to put more focus on relationships in the near term, right? Like, ⁓ cool. Would I rather have my CSM going to focus on the relationships right now or going to focus on doing a check in, like doing some tasks that they've been doing up until now, checking spreadsheet or doing something else. Like, no, I'd rather, like right now I'm almost like over indexing, like, hey, you need to be like so in tune with what that company is doing, what's their strategy, where they're going. Like you need to be like, essentially you need to tell almost like, tell me that you work for that company. Like it's like the way I'm thinking about it right now, where like, you need to be so ingrained that like we know what their next moves are. Jay Nathan: Yeah. Jeff Breunsbach: And like, again, I just don't think that CSMs have been in that place, right? You've been able to kind of get away with, we have a QPR. Let me just get some info. That's enough for us to appease ourselves and kind of move on. And I think like, you've actually got to lean like so far into like, you're almost like an employee of that company that you can tell me and recite all their information. Because that's the value that I have in that role right now. Like, like you said, like I can go figure out how to automate a task on the backend. can go, I can take our Fathom call recordings and I can go. Jay Nathan: Yeah. Jeff Breunsbach: basically do your follow up, do your call notes, do you know, like that stuff is becoming more real. Like it's almost like that guy was saying of like splitting the job description. You know, it's like, cool. I'm taking all the parts that I can automate. And so like, where can I not automate? And that is you spending time with our customer on the phone, like building a relationship, understanding who they are, and then really getting to the depths of their business. because a lot of like the other thing I'll mention too, is like a lot of these businesses that are private, right? It's hard for Jay Nathan: Yeah, yeah. Jeff Breunsbach: it's going to be hard for us to use AI to get this public information and say, okay, like I know their strategy. know where they're going. I know what products they're building next. Cool. I can go do that for HubSpot for they all have quarterly earnings calls. Like that's a good indicator of that for us, but like private companies, you don't have that level of information. And so like we need to go get it. And like, that's to me, like where the CSM is best positioned. And like, to me, you've got to lean harder into that right now. Jay Nathan: If you're carrying that cost already, then I totally agree, right? There's two things come to mind from your comments, and I don't totally disagree with you, but I think there's maybe a couple of different ways to think about it. Number one is. We're making the assumption here that, um, I lost my train of thought on number one. It was so good though. God, it was good. All right. But number two, let me, let me skip ahead to number two. number two, think about if you are coding a new application right now, or if you're even, you know, if you're building something simple and you instruct an agent to go, you know, build that or ask it what the best tool is, it's going to go out there and find, and you need to take payments. Okay, let's do it. It's probably going to go out and find Stripe. It's probably going to look at Stripe's API and it's going to hook that up and connect it for you. So I think that's agent to agent PLG, right? it doesn't have to be the, the end-to-end solution, but like that's going to happen before you ever get a chance to talk. people. So that might be just like the new marketing agent agent kind of purchases for a specific thing. Now, I do think the more complex the solution gets back to our matrix, right? Like how many functions does it cover? How many in the are going to be impacted by whatever solution we're talking about, whether that be CRM or, you know, zoom, ⁓ zoom recordings, it what are the competitors? I think way that it's deployed in the organization are going to be factors there. And the simpler the solution is, or even maybe a simpler part of the solution is, like, they're going to be internal ops teams vibe coding tools for their own teams. That is happening already, right? And I wouldn't even call it vibe coding because I think that undermines what's actually happening there. Like we've built a wrapper around chat GPT for our company. Jeff Breunsbach: Yeah. Jay Nathan: that encapsulates our customer and project data in a way that's safe, meaning it's not training the models on our customer data, but it's constraining conversations with Chad GPT down to that customer's context and projects and data. So we that tool together and when we ⁓ go and ask it to connect with some service that we need, it's gonna go out and find the best service. When I think about that, the, that's what I'm, there's just not ever going to be a chance for a human to even be in the loop on some of, on some of that from a demand standpoint. Now back to my first point, I do think what you said is true about CSMs and you if you're carrying them on the books, you're not ready to cut that team. Then by all means go get as much information, go get in bed. I think to the point of the article. deploy them as far as you can to make sure adoption of the product is happening. in the relationship side of things, there's a lot of fidelity lost in to do things way. You're making the bet that the customer even really wants to get on the phone with somebody at this point. And if they don't have value to add, then from a product or process perspective, then I would say it's like. Jeff Breunsbach: Yeah. Jay Nathan: three times more likely that they don't want to get on the phone with your CSM. You know what I mean? So it's not all about CSMs. And we say that all the time. It's about any facing post sale resource we put in front of the client. Jeff Breunsbach: Yeah. Yeah. Yeah. I think like the response that I would have to you right now is like to your support question the other day that didn't get answered and you threw a ticket over the wall. Would you have much rather had somebody just pick up the phone and call you a minute after you said you had a problem and fix it then? Jay Nathan: I would have rather they texted me not called, but yes, ⁓ actually, I did have an experience like that this week too, which is pretty cool. I could tell you about that. It was actually a good, a good support experience. Jeff Breunsbach: Sure. Sure. Yeah, But I mean, yeah, but I think like to your I think like the I think we're probably going down the same path. And I agree with you, which is trying to think of the right way to say this, maybe like, lose a lot of credibility with the customer and drips across the entire customer journey. Cause we get on a call and then that call doesn't actually end in the outcome they were hoping for because we have to take it back to our team. We have to go find the answer. We have to go do this. Right. Like, and I think that happens too often. And that happens, like you said, across teams, right. whether I'm putting an implementation person for them, CSM, support person, but like every time we get on a call and something's not achieved for the customer, like I think of it as almost like a chip where like the customer's like, okay, cool. Like you just wasted my time and took it. You know, I'm taking my chip from you. And like you said, like, I think that that's where you, and so to me, like, how do you marry those things up? Like you said, how do I get that? How do get that person like deployed so that like we can go accomplish the thing that we need to, if we're going to require the person's time on the other end and like that. Jay Nathan: Yeah. Yeah. Jeff Breunsbach: That to me is like the, I guess the wrapper of like the way to think about it. Jay Nathan: Yeah, yeah, yeah. Totally agree with that. Totally agree with it. Jeff Breunsbach: Okay, cool. right. This was a cool episode. So pretty fun. think talking through a couple of LinkedIn posts, you know, think about some of the AI pieces of how you're rolling it out to your team, you know, thinking is it team first, you let them kind of do it from ground up or top down. I think talking through a couple of the things that, you know, people are doing wrong right now, just automating bad processes, kind of immediately jumping to I think you ⁓ your kind two by two box, also just showcased building something, you know, pretty quickly with Claude that like anybody could do. ⁓ And like if you download that HTML, it's what I did the other day, like you can actually just go host it ⁓ on website, which think is cool. And then ⁓ last I think about some of these, you know, what's the new norm start looking like with AI in picture, some of these software what are some of the new metrics or ways that we would be looking at successful companies? Jay Nathan: Yeah, good episode. So the only question is, are we going to call it? So I have to brainstorm. Jeff Breunsbach: ⁓ I don't know. Yeah, we will. All right. We'll to you next week. All right. Jay Nathan: All right, See you.