Jeff Breunsbach: All right, welcome back to another episode of the Chief Customer Officer Podcast. It is Thursday, August twenty seventh. Jay, what's going on?

Jay Nathan: Not too much. just another day in paradise, man.

Jeff Breunsbach: Down in Charleston, no traveling this week.

Jay Nathan: No traveling this week. I was in Boston last week, which I'll tell you a little bit about. I I got to go to the CCO summit up there, which is pretty cool.

Jeff Breunsbach: cool. Yeah.

Jay Nathan: couple of observations. But yeah, glad glad to not be traveling this week, but I do need to be on the road more. I was also at a QBR, a sales QBR for the first couple of days of the week. And one of the big themes of that QBR for those folks was we gotta be out in the field with customers more. And

Jeff Breunsbach: Yeah.

Jay Nathan: so

Jeff Breunsbach: There's

Jay Nathan: That that's a theme.

Jeff Breunsbach: there's still something about kind of like collaboration in in person and I don't know, there's something about that that you you know that works. I kind of left my team off site thinking the same thing where it's just like it's it's even less about the moments you're spending like in the room together and more about like some of the small smaller moments, some of the s chit chat, some of the jokes that you make, like all those things end up, you know, there's

Jay Nathan: Yeah.

Jeff Breunsbach: something about that that helps you gel as a team better.

Jay Nathan: When you have a when you have a real connection with people, they're more likely to communicate back to you and you know,

Jeff Breunsbach: Yeah.

Jay Nathan: you're not just an email, right? You're not just a slide message, you're an actual human being with feelings and and a need to accomplish something. So yeah.

Jeff Breunsbach: Yeah. Well, there's there's like three things I want to talk to you about that I've built over the last week. I've just been like, you know, a shipping machine over here. And

Jay Nathan: You're crushing it.

Jeff Breunsbach: the first one is uncommon. So we launched our community out to the public. so if you're listening to this, go to uncommon.chiefcustomer officer.io. We have built a fully customized community solution for us, which has been really cool to play with. So I would say some of the things over the last week that I've been playing with is like vector databases. I had nothing, I had no idea what those things were and did some research and it helped me go through and set it up. and now we've essentially got our own vector database that's stood up that we're using for some of our community findings. And that's really to try and help us create almost like a I guess a more cross-pollinated community so that posts and members and tools that we use all And then newsletters and podcasts essentially like all become interconnected. and we can actually make inferences between those, which is like to me, I think part of the fabric. You don't have to, you know, you don't have to remember to always tag a certain thing for it to be pulled into another place. And that to me is like, you know, trying to not make humans remember everything every little thing that you have to do in the community and and almost like making the community work for it. so we ship that. I think it's been fun the first couple of days. We'll start. to to push that more and more. But even today I thought there was like, I don't know if you saw, but Ian, who's a member in our community, posted about agents and how are

Jay Nathan: Yeah.

Jeff Breunsbach: people how are people deploying agents safely and how are, you know, there it's the age whole question. Like he's he's kind of like, hey, I can deploy agents on my computer, on my own personal machine, but then like how are we doing this, you know, as a team? so I think we've talked about that a little bit here. And I think there's gonna be a good discussion in the community about that. So that's the first thing that we launched. Or that I launched over the last week, which has been pretty fun. The second thing is a training academy for my CSMs. so right now we have you know, again, early stage company. So I don't have a you know, we don't have a very robust onboarding experience. There's not a learning and development person that we have. I don't have enablement, you know, again, we're we're kind of a scarce amount of resources, and so I've got two new C I I just had two CSMs or a CSM and a solutions engineer start about a week or two weeks ago. I've got two more CSMs starting next week. And and I'm kind of in between this problem where like, you know, I'm onboarding people that I can't I don't really have the materials. And so I thought this was pretty slick. What I did is I took our API docs, I took our Slack, I took our Notion pages, took our website, took our current onboarding experience. We have some materials that we put together. And essentially I asked Claude to basically go build me a an LLM experience for my team and it came back with eight lessons. There's about three to five topics under each lesson. it's it follows what's called the plan. I don't really love the name, but the plan is essentially you should go through it in that in like in this series of of an order, which I thought was great because it basically took we have some market analysis that we've done. So it kind of tells you about the market that we serve, where we sit in that market. Who are the players? It kind of gives you that overview. Then it drills down into okay, what's what is junction? What's our solution? How are we presented in the market? And then it starts to get into like more of the kind of the products and some of the the nodes essentially that you can have in a product. talks about some of the you know the hierarchy of an account and an org and a team, you know, why does that matter? then it gets into our products and and how some of these things are delivered for our customers and the outcomes they're looking to achieve. So so yeah, I spun this up last week. My team's going through it right now to help. You know, I think the biggest thing I have to do is validate that the information is correct and that we're like training the right things. But I think for like a one-shot, you know, Claude, if you look at the experience, it's got basically it's got lessons, it it actually embeds videos, diagrams, it has note cards and flashcards, it did quizzes, and so it's like tried to create this, you know, experience of like, hey, learn something, do something. and then also what I'm connecting up right now is a way for almost for you to play with our API in that experience as well. So our teams can actually see the product kind of working as they're, you know, learning a lesson. So so I think there's a lot that I need to, you know, go massage around that. But at least like I sent it to, you know, a somebody on my team and like the first their first reaction was like, Okay, this is cool. Like, totally love that you just did this because you know, they were kind of going down the path of like learning stuff just by like Listening to Fathom Calls. that's the other thing is I I took Fathom Call transcripts and that's embedded into every lesson. So it it essentially lets you go watch it go lets you go watch something

Jay Nathan: that's cool. Yeah.

Jeff Breunsbach: live. so anyways, that's that's something I built. I'm curious if you've got thoughts around it. Maybe maybe you have an idea or two how I can make it better.

Jay Nathan: The only thing I was thinking is, and I've just started doing this with every I mean this is really I've started doing this with everything that I'm building right now for our team too. And I say I, we. but what is the learning loop? So you mentioned there are quizzes, right? one of the things that I that I really have leaned into over the past probably month is making sure that when I when I have when I build an agent to go do something. Let's say it's in cowork, right? Very simple. But then I always have a file in there that's called like lessons learned. And as I'm coaching it, if if if you I have one agent that's got a skill that basically the the skill has like it's it's a content type of agent. And what it does is that skill walks through all the different steps of the the content creation process, the content discovery and creation process. It actually has some cool stuff built in there too. I actually stole this from Alex Lieberman and his you probably heard that podcast on Clairvaux. I

Jeff Breunsbach: Yeah. Yeah.

Jay Nathan: created that same exact structure and agent. So it actually interviews me like like interviewers, like when I want to go create a piece of content, it interviews me. But the the cool thing is when it spits out the final version of the the piece of content that I'm starting with, then I go edit it myself. And then I give it the feedback. And then the last step of the process is that it takes that output and compares it against what it wrote. And then it logs everything that's like, okay, here's what he did differently from what I did. And it logs that as a lesson learned and then

Jeff Breunsbach: That's all.

Jay Nathan: applies it the next time. So I was just thinking while you were talking about the training thing, like how do you start to get that sort of continuous learning loop going with with the agents that you're creating? I think that's sort of the difference between, you know, building a task tool, a tool that just does a task versus something that learns and and can be continuously valuable as you

Jeff Breunsbach: Well,

Jay Nathan: keep working on it.

Jeff Breunsbach: it's funny you mentioned that. So so the the next place that I went and one of the other problems I'm trying to solve, and it'll kind of tie back to this to this original academy idea, is we have API documentation. That's probably the most robust thing that we have here. And then if you were to go look at some of the non-technical pieces that we have, you know, it's sparse. It's it's hit or miss. Did somebody write a guide for that? Is there we have a ton of Slack messages? I mean we Ad nauseum, one of the things that I am like

Jay Nathan: Yeah.

Jeff Breunsbach: drowning out about is like we cannot allow decisions to live in Slack because they just get lost. Then you can't find the thread. And then you know, again, we're spending time going back. Then you're asking Claude to find it, and Claude's like, Well, I have this thread from this state and this thread from this state, which one's real? Blah, blah, blah. So

Jay Nathan: Burning tokens.

Jeff Breunsbach: yes. So I've so I've got our solutions engineer and myself working on this project, and we're trying to I'm I'm trying to like move at lightning speed because We need an internal wiki or documentation. We need an, you know, some sort of LMS, some sort of content repository, I don't know, knowledge management system. I don't know. There's all these words that you can throw out there. But at the end of the day, I need to be able to essentially like capture decisions around our product, how it's used, best practices, all the trainings, you know, what are we rolling

Jay Nathan: Yeah.

Jeff Breunsbach: out? the SOPs for each department, right? You start thinking about this. and so we've been going down this path. And I just read Is it Andrea or Andrea? Andrea Caparthy, who's like the Andre Caparthy or

Jay Nathan: Andre Andre Carpathy, yeah.

Jeff Breunsbach: Pathy, sorry, wow, what a butcher of a name that I just did butcher job of a name. so he wrote back in April about how he's been building his own wikis and

Jay Nathan: Yeah.

Jeff Breunsbach: like his philosophy around how you could go do this and how this could become. Like a knowledge management system for like one of the use cases he says is like this could become a s a knowledge management system for a company. And it's all in a GitHub repository and markdown files. And you essentially then like you you actually don't touch the core files itself. You have the system touch them. And then you create these loops or these lints that, like, okay, like you said, like, okay, it needs to be learning itself. It needs to be like we need to be able to have. a queue of like, okay, these are updated. Are these accurate and verifiable and whatnot? And so so right now we're like in the midst of this week. Like we just have figured this out in the last like 24 hours. And so we're going to start building out this like internal wiki in GitHub, markdown files. but now it's like you said, now it's got it's going to have like three agents that are working on that thing continuously.

Jay Nathan: Yeah.

Jeff Breunsbach: so there's going to be kind of a a writer who's writing the original. pieces. There's going to be like a reviewer who's reviewing the writing style on the code and everything

Jay Nathan: Yep, yeah.

Jeff Breunsbach: else. And then there's going to be the third one that's running the loops of like, okay, who's asking questions in Slack? Is that is that captured in the in the wiki already? If not, let's flag that. it's going to be going through and also like verifying. And then there's going to be like verification around almost like each bullet point that you have. Is it, is it like truly verified? Is it like a you know, Is it like highly verified? We're like we we have it in the Slack thread from our product leader, but like he hasn't really confirmed, you know, like there's gonna be levels of this so that we can understand. but I guess like the reason why we went down this path and what I just found so interesting is and and maybe this is where you're going to, I find myself still designing things for humans instead

Jay Nathan: Mm-hmm.

Jeff Breunsbach: of thinking about the agents. And when I read his thing, his whole thesis was like, like, don't worry about the human, like, yes, you need I I can go create a UI that can easily like I'm gonna go, I'm going to go deploy a Vercell UI around this whole GitHub repository so my team can go find what they want. But he's like, the reason you do it in in this style and format is because then you have the agents running and operating on files that they're already used to running. And like that's the whole thing. Cause we were thinking about how do I push this into Notion and then like what Notion pages, how do I create multiple databases? How do I connect pages to each other? And like as I read his thing, I was like, my God, I'm like thinking about this all wrong because I'm trying to I'm trying to push it into a tool that I already have versus thinking agent first and like I need these agents running on files that they know about and it's in a GitHub. We can log again, we can create logs and systems and files and ways that we can make sure, okay, we know what's happening, we know what's being updated, we have a a queue

Jay Nathan: Yeah.

Jeff Breunsbach: of things that can be reviewed. but now, you know, I can go create a custom UI, which is even better probably for my team, and then Now I've got a system that's actually running in the background. So it's listening for Slacks, it's listening for emails, it's looking at support tickets and pylon and all of that stuff should almost be self editing. And again, the whole his whole point was like, and then you should not be touching the source files because you should have the agent touch the source files so that again, it's like it's it's a it's its own living breathing thing.

Jay Nathan: Yeah, it's truly autonomous. yeah, that absolutely. And the this is where like we've used the word ontology before. Like that's where that comes in. So what are the things that you're having agents go build from all the data? So it's you have to separate the the the data layer from the tooling itself.

Jeff Breunsbach: Yes.

Jay Nathan: to your point. And I think that's where that's where a lot of folks are getting tripped up because You have all these tool vendors out there. I can do this for you, I can do that for you. And but but what you're describing is is the right way to do it. I actually think though, the challenge with it is both the build and the maintenance of these things. It's sort of it's a little complex, to be honest with you. And I just think the average the average team is just not going to build this and probably shouldn't be building it. So I I do think, and it gets into maybe the discussion we'll get into later if we have time, but I do think there's a there's a product layer here, but the product layer is completely different than the typical SAS tools that we're used to using, not just because those are seat-based tools that you're assuming a user is going to edit the file. It's that, you know, these these tools need to be agentic from from the start. And so You know, once you decide what your ontology is, I and by the way, ontology is just a fancy word for like customers, employees. We have engagements in our ontology and at Balboa, right? Because we have engagements with our customers. it basically describes all of the business entities that this is where design thinking really comes in handy, right? It's like, okay, what do we do as a company? What are the things we interact with? What are the like contracts that we have? Like it's all like very simple once you start to you know, back out a little bit. But then you have agents that are purpose built

Jeff Breunsbach: Maybe that we can start to build.

Jay Nathan: to go maintain that. Now imagine if you had a CRM system that was built that way. That's where this is headed, right? Because the the the CRM system is really just like we have agents that are running over everything that you just described, Slack, email, meetings, fathom recordings, all the content we can find. We have agents that are running over all that data, just like you described. And instead of producing knowledge base articles about the product, they're producing essentially the the equivalent of a knowledge base article about a customer and about a contact at a customer. And that's that's our continuous learning loop. And then also there's another thing, like we have best practices that we share with our clients about how to implement Pendo. Well, guess what? Those are all being harvested out of those same calls. So that's a different agent that's pulling that out.

Jeff Breunsbach: Yeah.

Jay Nathan: And there are open source tools. Actually, Obsidian is not open source, just to be clear. But

Jeff Breunsbach: Yeah.

Jay Nathan: you can take a tool like Obsidian, which is pretty much it's very cheap or almost free to use, and you can actually just use it on top of a file structure that's sitting on your server or your or your laptop, and you have an interface, right? So you don't have to put anything into Notion. Maybe Notion, you can do this with Notion too. I don't know. But anyway, so what you're saying. Is absolutely in my mind the right way to be thinking about this stuff and I mean, Carpathy's like the the he's the alpha geek on all this.

Jeff Breunsbach: Yeah. And this is so this is just like a I just wireframe this. This isn't real, but this is I mean, in in my mind this started to become real if you can see my screen, right? Like this is a

Jay Nathan: Yeah, yeah, yeah.

Jeff Breunsbach: wiki. It it it back it again, it backlinks because it knows all the other content that this touches. it looks at the sources, you know, it tells you what's shipped were versus what's proposed. It kind of gives you like a local knowledge graph that you can see, right? And then like

Jay Nathan: Yep.

Jeff Breunsbach: to your point, what I can then start doing is going through okay, what's the index look like? Here's the log of everything that's happening.

Jay Nathan: Yes.

Jeff Breunsbach: Lint. So here's the con here's the conflicts. Here's the unsourced claims. You know, here's all this stuff exists. This is an example of like how it would, you know, work in Slack. And then like how does this get back into the system? And then like it kind of gives like this wiki and gest query lint. But like,

Jay Nathan: Good.

Jeff Breunsbach: you know, like to me, again, like this this to me, I guess like again, more of the power at your fingertips. And and I actually think that this is like a relatively simple solution because actually, again, like you I a bunch of the stuff already exists in Notion pages and everything else. And so like we can basically go almost like get this up and running this week and then start to review files and have like, you know, again, like there's gonna be some review from our team and from our from other teams. So like there's work that has to go into

Jay Nathan: Yeah.

Jeff Breunsbach: this. But I guess like the work that I don't want to happen is us to have to go figure out which tool do I go by, how do I then fit, you know, what what's the data layer? I have to create all these, like you said, I have to go create the The data model that I want to make sure I can, you know, link all these things together and everything else. Like that I don't want us to be spending time on. I'd rather us spending time on the actual like verification of like, is this the right information? Is it accurate? Are we

Jay Nathan: Yeah.

Jeff Breunsbach: is this truly a best practice we should share with our team, you know, with our customers? So, anyways, that's the third. So then what got me thinking is like the academy, you could then think about all the things that you could power from an internal wiki. So I could I could power my academy. I could create most likely. right sections for the right teams, you know, for people who, you know, what what parts of the product do you need to know if you're a support person versus CS versus a salesperson, you could probably start to customize based on that wiki. Then

Jay Nathan: Yep.

Jeff Breunsbach: then the other one is like customer facing docs. So like from that, can I start to create a custom can I have a skill that's essentially combing that stuff again, almost like creating a a a secondary tree of like customer facing docs that are MD files.

Jay Nathan: Yep.

Jeff Breunsbach: And then trying to figure out, okay, how should I present this information to our customer? What's relevant to them? and so like to me, I don't know, you have this internal wiki that is ever growing. It's not stale because it's listening versus us having to update it ourselves. And if it's not stale and it's updating, then that means our customer docs are are updating. I can be, you know, I can I can be looking at our API docs. You know, when do our API docs update? Can I have a webhook that allows us to know when that happens? Like, so there's just I guess so many things that to me, this internal wiki could power for you that you know it doesn't totally eliminate the problem, but it just makes it so much easier, especially for a small team. Think about if I have to go create customer-facing documentation. That's a that's normally somebody's job. They have to go write that. Okay, I have to build the the articles. I have to go, but now if I have this thing powering it, maybe somebody's job shifts from totally writing everything and and looking at these pages to building the system. Yeah, building the system.

Jay Nathan: Viewing it. Yeah, yeah, yeah.

Jeff Breunsbach: Building the training, it you know, reviewing the stuff. So it it also, I guess going through this process and looking at a bunch of things that we're building right now, I guess continues to enlighten me. Like jobs aren't going away, it's just that they're shifting, right? rather than writing support articles, I'm now creating the system for writing support articles. I've got to train it. I have to think about the loops. I have to think about new sources. I have to think about, you know, which one takes precedent and you know, how we're gonna make some of these decisions about which one's accurate versus not. Like, but it just changes. And you still need somebody to do that. and if I can start this now, the person who gets to pick this up now has I'm such a head start to, you know, helping us do this work.

Jay Nathan: Yeah, absolutely. Yeah. I mean, I I don't think there's any evidence anywhere on the planet that jobs are going away because of AI. I actually have to talk to my one one of my kids about this all the time. 'Cause, you know, she is i let's just say she's in the anti AI camp. And I'm like, Okay, you've been watching a little too much TikTok, a little bit too much Instagram here. college student, yeah.

Jeff Breunsbach: Your out well, your algorithm is feeding you.

Jay Nathan: Yeah, exactly. I'm like, well, you know how your dad earns a living, right? To to pay for all your all your stuff. But no, I I think I think I mean what you what you're what you're describing is like it's spot on, man. and you know, it does it actually creates more work for us. I think what what you're you're finding is that it creates more work. There there are a lot there's a lot of you know, we've talked a little bit about this, but if you zoom out a little bit, that part of the problem we have with AI is a messaging problem. In that people that are running these shops, particularly Anthropic and Chad GPT, over the past three years have preached doom and gloom. And, you know, they they think they're creating something that's going to, you know, basically end civilization, but it's not true. Like we're not going to have universal basic income because there's nobody that's going to sit on their laurels. They're going to look at this stuff and figure out, my gosh, if I can automate all this stuff and and Have it running autonomously, then what else could my people be doing? We're gonna keep innovating and creating new products and new businesses and new solutions. I will say, all of this automation, I think we talked about this before, but it's pushing the problems around a little bit. We see this every day now, both with our partner who is shipping more code and more more software than they've ever shipped by far, right? I think that on their quarterly on their quarterly product update. webinar, they said, Hey, in the first half of the year we shipped more than we did in the whole of last year, right? And

Jeff Breunsbach: wow. Yeah.

Jay Nathan: by the way, it's like exponentially accelerating like that. And it's gonna continue. But the the problem is not shipping code anymore. That used to be the bottleneck and software companies, you know, have to have to wait on engineering to get something done. And engineering is, you know, rightfully probably slow and methodical and d doing things You know, correct. And of course, like you but that's where the bottleneck was. I mean, you were typing every line of code by hand back then. That constraints started to go away. Now

Jeff Breunsbach: Yeah.

Jay Nathan: everybody else is scrambling to keep up. Sales enablement. The sales teams cannot stay enabled on all this stuff. The customer success teams, God forbid, the customers, right? If the teams inside the company can't even stay up to speed on what's shipping in these products, how on earth? Are the customers staying up to because the customer doesn't think about you every day? The customer

Jeff Breunsbach: Yeah.

Jay Nathan: doesn't probably even use you every day, right? So, like this is where the bottleneck is now. And I think our our big challenge in whether whether you're in a go-to-market team or you're part of the customer success organization, or they're both one, whatever it is, that's where the problem resides. Product marketing is under pressure.

Jeff Breunsbach: Yeah.

Jay Nathan: Like every everybody is under pressure now, it seems, except for. Except for maybe engineering and product.

Jeff Breunsbach: Well it's it's it also I I think illuminates for me too, or like the the route that you're talking about, right, is is also I think it becomes even har or it becomes even more important for you to for companies to be asking, are we like shipping the right things? Right. Like you can

Jay Nathan: Yes.

Jeff Breunsbach: you can now just ship anything. And so now

Jay Nathan: Yep.

Jeff Breunsbach: again, like you said, right? Like, we can go solve so many things so many more things for customers. But like you said, then in turn, if you turn around customer actually is y it's almost like it's almost like reverse ROI where you're like, well actually you're just bogging them down more, which means they can't actually get more ROI to your product, which means then like they're actually more inclined to like, my gosh, I can't tell this story internally. So therefore like I actually

Jay Nathan: Right.

Jeff Breunsbach: can't afford your product anymore. And but you're telling me you shipped, you know, more features than ever. And it's like, but it's the you know, it's like the the problem of like, okay, are any of those features actually helping me towards my outcome? and is it, you know, is it really Yeah.

Jay Nathan: That's right. It's less about features than it ever has been. Yeah. I imagine like a like the visual I get in my head about this is, you know, a speed boat, right? And if a speedboat is going across the ocean and it's going really, really fast and it's one degree off

Jeff Breunsbach: Yeah.

Jay Nathan: for when it starts from where it should be going, well, it's gonna end up in a different continent when it gets across the Atlantic Ocean, right? It's gonna

Jeff Breunsbach: Yeah.

Jay Nathan: end up in Africa instead of Europe.

Jeff Breunsbach: Yeah.

Jay Nathan: Maybe that's a little extreme. But like the point being, like it's really easy to go in the wrong direction really, really fast now. And so actually I was we have a this digital experience roundtable for product folks on Thursday. and I was I was prepping with it with an old colleague of mine, and we were talking about this very thing, right? It's it's easier than ever to build the wrong things and to do it really, really fast. And so you you need you really need two things. One is discernment. About what you should be spending your time on, because even though it's faster, it's not free. And then number two is you need a very quick learning loop for the organization to say, okay, like product market fit, where are we this week? Are we headed in the right direction? What is the platform we're building toward? Do customers want that? is this valuable as a cohesive set of capabilities? And that answer is changing so quickly across almost every domain. So

Jeff Breunsbach: Yeah.

Jay Nathan: yeah, I think product management is more important than ever as a discipline, as a result.

Jeff Breunsbach: Cool. Yeah. well I know we've got roughly we'll say like seven, eight, nine minutes or something, but it sounds like you had a good round table up in Boston and had a couple of of things that you wanted to chat about, maybe some observations from kind of what you're seeing in the market.

Jay Nathan: Yeah. so I went to this CCO summit. There was a bunch of other, you know, functional summits going on at the same time, which is cool. I don't know if you knew that about these things, but

Jeff Breunsbach: No.

Jay Nathan: it's I think the Customer Success Collective, not trying to promote them, that's just who it was. they also have product, yeah. Yeah, yeah. So they had CROs,

Jeff Breunsbach: that and they have other collectives, they have like a product collective and a and so they try to launch like one event and then have leads like multiple rooms going. Got it. That's kinda cool.

Jay Nathan: they had chief product officers, they had Chief people officers there. and then what what else? I don't think they had a finance track. Anyway, so it was mostly go-to-market facing teams. but the, you know, the content was okay. there were some notable standout presentations. Our friend Christy was the MC of this thing.

Jeff Breunsbach: cool.

Jay Nathan: but you know, like all these things, it's a very local event, right? So this was in Boston and it was good networking with the Boston crowd. Like you you go to the the events free. For attendees. So you just have to sort of apply to get in to make sure you're the right level and the right role to be relevant to the conversations because that's where the real value is, is in the networking. what the the one observation I had was that there were it was literally only one person presented all day that I saw at least who mentioned anything about true PL and balance sheet metrics. And You know, everybody of course throws around net retention this, gross retention that. Half the time I don't believe people know what they're talking about when they say those words because they actually don't. not to be rude about it, but I think there's a lot of surface level kind of stuff shared in those those kind of conferences and rooms. but it just got me thinking about our so for those who don't know, if you're new to listening to Jeff. in and me and our stuff, but we built a course for financial intelligence a several years ago now. and it's still relevant, because it teaches you about what a PL statement is, what cost of goods sold are, what gross profit is, operating profit, all that, all the basics, right? That like if you're anything like me, you learned it in college just enough to pass the the classes and then you didn't really learn it until you got into the real world where which is where I actually learned it. But it got me thinking about like this idea of getting back to the basics on that stuff again. It's still it's probably more important than ever because there's also, and this is what it led me to, there's all this conversation going on on X and LinkedIn about pricing models and how SaaS is dead and the seat licensing model is going away, which is probably true to some degree. But what replaces it? Consumption models or outcome based pricing? But those two models have some very important and nuanced financial implications to the business. They are fundamentally different companies, right? You you mix that with the additional costs that we've inherited from AI inference for solutions like you just talked about, and the whole game is different. And I think you actually need to know this stuff. So we don't have enough time to go into it today, but maybe we'll sort of tease it out. And we can have a longer discussion next week about like let's just go back to the basics, walk through like some of the core metrics from the PNL and then talk about how those metrics are now changing that now that we have different pricing models being talked about and different cost implications relative to the technology underneath it. I think it'd be a fun

Jeff Breunsbach: Yeah.

Jay Nathan: conversation.

Jeff Breunsbach: Let's do it. Well, one thing I think we can't maybe like we can just like talk about real quick is a question that I was having. So I was doing some research around, you know, this topic ever since you mentioned it. And one of the things that I was walking away with was because of because of this new AI consumption model, right? Because of like getting to outcomes, is there actually more of a pressure on CS? To get out of the features and adopting the product and worrying about, you know, how many times you log in and all these things. Like, is that is there actually like far less of an emphasis on that because again, like your customer now is getting to a point where this stuff needs to be embedded into key workflows. There's probably agent, there's agents surrounding this now. There's there's kind of all these things that now lead towards, okay. What did this actually do? Like the cause it it can get so close to the outcome itself just by running through some of these processes and workflows that like it, I guess like to me, it's really, I guess, forcing CS teams to look at this and say, okay, you know what? We've probably been, you know, we've we've probably just been glossing over a lot of things and putting, you know, and and putting kind of lipstick on a pig saying, like let's talk about this new feature release, let's talk about this versus again, kind of the the object of what the feature release is doing for you. It's leading to this outcome and that's what we really should be talking about versus like the feature itself. Like I don't know if it if I don't know. That just seems like to me like a a thing that's going to force teams to start to again kind of change orientation and say, Okay, we're actually talking about the end goal here, not the not the feature itself, which I think

Jay Nathan: Not the intermediate outcome. So yeah, interesting topic because one of the things that we're actually training, getting ready to train our whole team on, I should say retrain our whole team on, because this is something that is just a recurring thing. We always have to come back to it. Is how do you how do you take what you're doing in in the product we deploy, which happens to be Pendo, product analytics, right? Usage analytics, how do you take that and map it to Ultimately to a KPI, a business outcome, a business measure like revenue retention or like cost of goods sold and gross profit, right? so the very simple framework that I have for this is lagging indicator, leading indicate, sorry, leading indicator, lagging indicator, KPI. So what we are going to coach our team on, I'm actually in the process of building this, and I'll share whatever training we create here with everybody because I think to your point. This is the job, right? To some degree. Like, or it's it's part of the job for sure. And it's part of the job if you're in sales too. identify the lagging indicators. What is like for us and pending? Okay, great. The product usage analytics, what's happening with the the features, what's happening with the consumption of the product? Great, but that doesn't tell me anything if I'm an executive. What what needle is that moving? What Cohort of users are getting more value? What is the value they're getting from that micro outcome in the product? Like what is the lead, what is the lagging indicator to that level? And then how does that lagging indicator impact the business outcome, which is measured by a KPI on the PL, on the balance sheet, on the cash flow

Jeff Breunsbach: Yeah.

Jay Nathan: statement, somewhere that a CFO could look at it and say, okay, I understand the mapping now. And the interesting thing is, and then I'll leave it at this. You have to make the the reason that this is so hard for people is because there's not always a direct correlation between those three things. You have to, and this is where strategy mindset comes in. You have to make a leap of faith that if we do X, Y is going to happen. And we believe, not we know, but we believe that it's going to impact this KPI.

Jeff Breunsbach: Yeah.

Jay Nathan: And then your job as a leader is to. vet your belief against what the rest of your executive team believes, what the rest of your team believes, what the customers believe, right? And say, okay, do we all agree that that's probably that there's a going to be some kind of correlation there. So that's where like the book Strategy Rules comes in. Like one of my favorite books, right, is actually not strategy rules. the root good

Jeff Breunsbach: Good strategy, bad strategy.

Jay Nathan: strategy, bad strategy, yes. So I don't know that that whole this whole body of Stuff just you know, it always gets me fired up. So

Jeff Breunsbach: Well, the other thing that this like you just alluded to, right? Like businesses are built on financial models. And there's a a whole aspect of like CFOs and the finance departments and the the pl the business planning side, right, is all thinking in terms of horizons in the future. They're trying to understand bets that you're making and assumptions and models that they're building. And to your point, like I think sometimes we get too baked into okay, well, what if my assumption is wrong? Or what if my what like, I'm going to say this and like what happens if it's wrong? And like in essence, though, that's I think that has like shielded people from again, like you said, trying to connect these dots and trying to say, okay, this is our best guess estimation. This is how we think about it. This is how the mapping goes, right? Cause again, like you just need a model or an assumption and that can, you know, that can take you as you start planning. And then you need to try and correct that as you go. But like if you don't have anything to react to and you don't do anything, then your teams are doing the they're they're doing the planning for you. It's just happening in the background. Then they're gonna give you a number, then you're going to complain about it. And then you're gonna say, How am I possibly going to do that? Right. Like it just creates a cyclical thing. And so like you need to go drive this proactively, like you said. Okay, we I know we're changing the business model. It's gonna be AI, AI consumption. How do I map all the things we're doing to the right KPIs?

Jay Nathan: Yep. That's a all right, we can have a longer discussion about it next week, but that should be a that should be a fun one.

Jeff Breunsbach: Cool. All right. Let's do it. All right. I'll see you next week. This was fun. We'll do it again. All

Jay Nathan: All right, talk to you soon.

Jeff Breunsbach: right. Later.