Latest / The Chief Customer Officer Podcast / EP015: Why CS Teams Are Building Their Own Stack

Transcript

EP015: Why CS Teams Are Building Their Own Stack · The Chief Customer Officer Podcast

• Jeff Breunsbach: All right, welcome back to another episode of ChiefCustomerOfficer.io. This will be coming out to you on Thursday, April 30th. End of the month recording, I love that. ⁓ Jay, how's it going? Sounds like you're up in Raleigh. Looks like you're in a lovely phone booth ⁓ right now. Jay Nathan: I how's the sound quality? Is it nice and tight? I forgot my microphone, so you got my, my AirPods, but ⁓ yeah. Jeff Breunsbach: Yeah, it's good. Hey, I'll tell you with AirPods and then a little bit of like Riverside Magic, would say our podcast sounds pretty good in terms of like removing all the extra noise. Jay Nathan: Yeah, yeah, yeah. Yeah, so up here in Raleigh for sort of like end of quarter and some partnership review activities here with Pindo and. Yeah, end of quarter for a lot of people right now, end of April, right? So busy time. Jeff Breunsbach: Yeah. That's cool. Yeah, we've got a try to think I've got like seven or eight weeks until ⁓ our baby arrives. So I'm in a little bit of a actually it's I think it's closer to six or seven weeks. I'm like, you know, I'm in the I'm in the last minute panic, you know, having a third kid. So we'll see. We have a lot of house projects that we're trying to finish and not a lot of weekends. And then also the weekend start stacking up when your kids have things I'm like starting to realize like my three year old has stuff to do. He has birthday parties. He's got t ball. Jay Nathan: ⁓ my gosh. Jeff Breunsbach: My birthday is this weekend. like all of sudden we're like, oh, we looked and it's like, oh, we have like four weekends already taken. So like, when are we going to do this? You know, might have to stay up nights and weekends. Yeah. But, um, yeah. Jay Nathan: Crazy. Well, just strap in, bud. You know what's crazy to me? I've never met Elliot. And I saw pictures of him the other day and I was like, oh my God, the kid's like a walking, talking human now and I've still never met him. Yeah, anyway, we need to solve that. Jeff Breunsbach: Is that true? That's crazy. Yeah, we'll have to. Yeah, yeah, yeah. He's like, I think he's at the best age, which is like just over one years old and he's starting to understand you. And like, it's pretty fun to watch him start to comprehend. And my favorite is he can't fully say words. And so you just have to interpret. so, you you'll be, we're trying to teach him thank you when you like give him something. And so instead of saying like, thank you, he's like, thank you. And he's just like these like big, you know, these like big transitions of words. So. This is gonna play great in a business audience who's listening to this first, know, five minutes of this episode. yeah, he's been super cute. well, before we jump in, I had a quick story. I'm curious if you have an opinion about this, but I've got a friend who was just telling me about, you know, they're relatively small company. I wouldn't say they're startup, they're kind of probably series A, I think, series A or series B. And his CEO is trying to get out a marketing video, like an asset that Jay Nathan: That's awesome man. Love it. Jeff Breunsbach: like they can use that's like a kind of a brand asset that's got, you know, messaging about them. It's got all their kind of brand colors and feel and all this stuff and kind of gets the customers hyped up and gets prospects hyped up. It's something that they can use as collateral. And so this buddy was telling me that the CEO kind of went around to some agencies and was asking, Hey, you know, what, what can I get? How much is this going to cost? And I mean, he was getting some, I mean, you know, like, He was getting some erroneous quotes like, you know, $500,000 here, $750,000 there in terms of like what he's asking for, like multi-minute video that's like well produced and like, you know, has a couple of different things. And so my buddy just said over the, over this past weekend, he just said, screw it. I'm going to go use AI myself. And he said with about a thousand dollars worth of prompts and like extra credits and all this stuff, he like weaved together a video and he's like, the video is probably 70 % there. of like what I want. And he's like, now I'm just going to ask my team to spend the like my internal team spend the rest of the 30 % like making it better. But the you he felt like that was a much better use of time and energy. So he said the CEO went back and said, Hey, what if you guys use AI, how much would the quote come down? And all these agencies came back and kind of said, Oh, maybe like $100,000 less, you know, there's maybe 10 % or 10 to 20 % of like, kind of like AI built into this and And so just curious if you think that that is, you know, are we more par for the course on those happening, on this type of thing happening more and more. Jay Nathan: I think, mean, yeah, the whole idea here is disruption and self-serve things that we could not do ourselves before, right? Like you asked me to walk up to, I don't know, Final Cut Pro, it's not Final Cut Pro, but like whatever the video editing software is, and I'm lost, right? And I'm not gonna put the time in to learn that. But when it's a prompt away, I am. So I have an interesting, so that was the CEO that did that? Yeah, yeah, yeah. Jeff Breunsbach: Yeah. CEO did that. I mean, there are about 100, maybe there are 100, 150 people. like their marketing team, said, was only a handful. It's two to three people or something like that. And so he's like, and yeah, yep. Jay Nathan: Yeah, right. They don't have that skill set on staff, right? Like to do the high quality video production. Here's the thing too, like for something like that, 90, 85 % quality is actually probably good enough for that, right? ⁓ And so the incredible thing is, and I think it gets back to what we talk about a lot on this show is, you know, what Claude Code does is gives a business Jeff Breunsbach: Yeah. Jay Nathan: person, the ability to generate exactly what they need. Like we were just talking about it and I think we talked about it on the last podcast, but what I built for my sales pipeline dashboard and on top of HubSpot is exactly what I need. No more, no less. And it's very simple, very straightforward. And I built it myself. I run it locally on my laptop, which is sort of a problem. But like I did that because I know what I need to. see myself on a day to day basis and what I need to share with my investors, what I need to share with my team, and it's all in one place. Same thing with video production. Same thing with content writing. Same thing with your website. By the way, I built our website the same way that I built that dashboard, right? Prompt by prompt by prompt, I knew exactly how I wanted to convey. And so there's nothing lost in translation, right? Same thing with that video. The CEO of most companies, whether they... Jeff Breunsbach: Yeah. Jay Nathan: admit it or not, they own marketing. They own the brand. They own the messaging of the brand. And so to be able to go directly and convey that is so powerful. Jeff Breunsbach: And that was like a big part of what he kind of asked the CEO, why'd you, you know, why'd you feel so compelled to do it? And he's like, I write all this material for you guys internally. I write all this material for custom. You know, I write all this stuff already. And he's like, why not just put that into a Google drive folder or like something that Claude can access. He's like, we have all of our brand templates. So he's like, literally I have all this stuff together. It's just that like Claude, and I'm not sure if he used Claude or not, but like, you know, whatever AI tool to use like. just brought it all together and he was able to get exactly what he wanted. And like you said, he was, I think you've mentioned this before too, right? But the idea of like being able to go kind of with a kind of a fine tooth comb and say, yeah, like, okay, here at this moment, I don't like this. I like this, or, you know, try and look, you know, change this out, swap this year, like the ability for you to actually go through and do that versus, you know, think about, I mean, I just think about the rounds of edits that, know, it would take now. Okay. Hey, take this document, go markup edits. Okay, great. I'm going to markup edits. Now I have to send it back to you. You're going to do it like, think about all this, you know, kind of back and forth that you essentially eliminate. ⁓ So I thought it was a funny example just in passing. Jay Nathan: Well, that's a really good point. That's a really good point. So think about the, so when you, when you have a team of people working on something, the communication problem, and I think everybody's probably seen the chart here of like the nodes on a, like a chart where you have two people, that's one line of communication. You have three people, that's three lines of communication. Then you have four people, then that all of a sudden it's six lines of communication. It's exponential. right, and it grows exponentially. ⁓ As you add one more person, there's a book ⁓ called the Mythical Man-Month. It's just sort of like the quintessential engineering book. And basically the thesis of the book is adding people to a project doesn't speed it up, it slows it down. And it's because of the communication. It's not because they're not capable, it's because of the work. So I think the superpower that these tools are giving us, whether you're an engineering team or marketing team or anything else, is to just reduce the number of people in the middle of a specific problem. And by the way, think about, I always go back to this, but think about all the calls that that CEO has been on where he has explained the mission and the vision of the company and what they do and how they do it to investors, to employees, to recruits, to customers. Jeff Breunsbach: Yeah, I think that's true. Jay Nathan: and partners, like he's said that story a thousand times. So why not go grab all those transcripts and use it to craft exactly the best of what that story should be, right? I mean, that's the opportunity in my mind is the words we say every day are, that's actually the work that we're doing. ⁓ Anyway, yeah. Jeff Breunsbach: Yeah. I thought it was a fun example to start with though, of like a real, yeah. Jay Nathan: So cool, yeah. We recently redid our branding, Jeff, and I think I've talked about this before, but when I first started working with our guy who's doing that, he's actually a good friend as well, and he's really good at what he does, but I was like, Chris, are you using AI to do this? Because I'm not paying you to design Pixel Perfect and write lines of code, right, when you build our website. And I think it was a little bit of a wake-up call for him. Jeff Breunsbach: Thank Yeah. Jay Nathan: But he's doing it right. Like he now uses these tools, but he still handcrafts the output, right? But he's way faster. He's like, I'm like four or five times faster now. It's like, great. Like I'm to pay you the same no matter what, because I'm paying you for like what you bring to the team. I want you engaging with my team. I want you really like pulling out what the brand is from us. ⁓ I want you crafting pieces of it, right? But I don't want you tinkering. Jeff Breunsbach: Yeah. The expertise, yeah. Yeah. Jay Nathan: with an image and I don't want you tinkering with WordPress. That's why we decided to build our website from scratch rather than use WordPress. It's just like, dude, that's way too slow for 2026. So. Jeff Breunsbach: Yeah. Well, I've got an example to share right off the jump with you. So, you know, we'll, think we'll have to find and weave a story back ⁓ about how to make this applicable for CS. But, you know, I woke up this morning and I was pretty frustrated because I'm like, I've been writing content. I think we've both been writing content since 2018, roughly 2017, somewhere around there. So we have, we have oodles of stuff out there already. Jay Nathan: Yep. Jeff Breunsbach: We've got a website now. We have a newsletter. We have all these things and some somehow I still can't consistently post when I think it's like one of the most valuable things I could be doing for my career for my business for not only for Chief Customer Officer, but also for Junction like I've seen how it's actually helped our recruiting pipeline how it's helped our you know could bring in like business opportunities and whatnot, so ⁓ So I just woke up this morning and just sat down. literally, think I've spent 35 minutes on this so far and I'm not trying to like exaggerate. Like this is literally just me prompting and Claude cowork for 35 minutes to try and get something that I feel like is pretty close. And so, Hey, I think you can probably help me. Like, you know, what's, what's something that you already can see that's like, Hey, you could do this better or like we could change this and then B I actually need to ask you like, how can I host this? I actually, or maybe I have to ask Claude, but like, I need to figure out this is just like right now running off of cowork and like a preview file. But like I need to figure out how I can actually like locally host this or put it on a GitHub or some sort of URL that I can like update. So that's like my end goal as well. But let me walk you through like the, so the idea is essentially like you have a week. ⁓ I want to be posting nine total times. I've got 40 ideas in the backlog and I have zero in draft right now. ⁓ And then this becomes kind of editable, configurable. So I can drag and drop these posts and move them around. So like, Hey, I don't want to do this post on Tuesday. I want to do it on Wednesday. I can move it. The concepts of the post themselves. So if you kind of look where it says anchor tactical opinion, curator distillation, I actually had it. downloaded my entire LinkedIn profile and I had it go through and I said, what are the posts that do the best and how should I basically build those style of posts going forward? And so this is like what it came back with was essentially like, There's an anchor. I want to be showing people how to use AI in real life. That is like my anchor. I should be doing that on Mondays. Tuesdays, I should be doing a tactical like, you know, here's a prompt. Here's something that you can use. Here's something that you can take away. Opinion is Wednesday. So like what's something that I feel strongly about curator? Hey, I saw somebody build something cool. Here's like, let me just share it with you. ⁓ Friday is like a distillation from the podcast, pull a clip, something right. We've got, then Saturdays and Sundays, I've actually been pretty inconsistent, but I think I'm going to start to highlight like companies and like a strategy. some sort of essay. And then I just have simple buttons down here. Is it drafted? Is it ready? Is it posted? I just want to click that thing and say, are these things done? And then if you notice this little draft button, it'll take this small prompt that I put in there. And I've already built a LinkedIn cloud skill that is based off of my top performing posts. And so if I click this draft button, it'll go draft a post based on my LinkedIn skill and how I want to write stuff. And then it's based off this little mini prompts that I put in here. Um, it'll go draft, uh, the post based off of that. So again, like, I think the thing for me that I've noticed over my lifetime of like building content is like, everything has felt too stringent, like, I'm locked into this. Like I've got to do this. But like, I, I, so I'm trying to stay somewhat locked into a certain process or like certain style of posts, but I'm allowing myself to move those posts around. I'm allowing myself to actually just generate the prompt based off of like a little. Jay Nathan: ⁓ cool. Okay. Jeff Breunsbach: a little box here. And then the last thing I'll show you is just down here is like simple, like this is literally across these categories. Like ⁓ it's just a backlog of ideas and it's just one simple line. Jay Nathan: And this is from your, this is from your history of posts. Jeff Breunsbach: Yep, history of posts. And then I also started, I don't think it's updated yet, but I've also asked it to go comb our podcast feed and comb our newsletters. ⁓ And then I'm also gonna probably figure out a way to try and go get old, ⁓ like gang-rotain stuff that we've done ⁓ too, and try and like add that to the backlog. So anyways, what's your initial thought? ⁓ know, yay-nay. Jay Nathan: So cool. That's why this was not very, very, very cool. What I was thinking is, yeah, like you were saying, like, how could you go pull more context into this to give you more forward facing ideas? So I think of, I use read wise to save articles. It's funny. I use it and I can't tell you the last, like I just went back and looked up that one that we were talking about, but I don't really actually go back to them all that often, but I know I have them saved if I really want to get back to them. Jeff Breunsbach: ⁓ that's a good idea. Yeah. Jay Nathan: And every once in a while I'll scroll through, but I'd like to be able to use that content, have something go rip all that content out and sort of comb it and then generate some stubs for me. So like what are the good sources of ⁓ things that trigger ideas for you during the day so that you can just store them away and incorporate them and let the agent go do that for you. By the way, I would call this your content agent, right? Your writing agent. ⁓ So that's really cool. One of the things I really like about this is that you're, so are you writing the whole post in the, in the app here or there's the prompt. I see that, but like what happens when you actually want to go write the post. Jeff Breunsbach: ⁓ so it'll, so right now what it does is it pulls up this little, ⁓ box and then I can post in, I can, so basically it's going to take my initial prompt that I have here. ⁓ it's going to take, other contexts. Like you said, if I want to throw in links or things that I've seen or other stuff about this certain thing, and then I'll say copy draft prompts, and then it'll essentially copy that entire thing. And then I can go plug that into Claude. So there's probably an automation step there. There's probably a. ⁓ Jay Nathan: Yeah. ⁓ okay. Yeah. Yep. Jeff Breunsbach: there's probably a way that I could tell it like, when I press this button, I don't want to have to go back in and ⁓ like copy and paste it. So it's like a ⁓ improvement I need to make right off the jump probably. Jay Nathan: That's where you would, so the difference between cowork and Claude code is that in cowork you can generate a file. This is static, right? So you took all that input and you generated a static output. And just so people who aren't watching the video here, which I don't know if we're going to publish the video or not, it's basically an HTML page that we're looking at and it's got different cards for the different days of the week. And then you can sort of, and then it's got a prompt associated with each day of the week that you can click it and then. take the output from that, which generates a prompt that Jeff paces into Claude that uses his LinkedIn writing skill and then ⁓ generates the prompt. then, now in theory, you could use cowork to drive all that, right? Because cowork can interact with different files on your desktop. So you could actually automate a lot of that. You can use, I think the schedules that are now available in cowork to do it. you could make... the hub for that. Or if you wanted to just have it all in one place and be able to manually interact with it, ⁓ then that's when you'd host it somewhere else, right? Because you do need it. You would want a database behind this ideally to sort of track everything and. Jeff Breunsbach: Yeah. Yeah, because that was like the next thing I was also trying to prompt it about was like, okay, I'd love to come back in. LinkedIn doesn't have like an open source or like a open API. So ⁓ I need a way to get some of the data back and like tell it how it's performing, how it's doing. And so like to your point, like I need some way I think to come back and also look at, okay, I posted this one on Monday. ⁓ I'll probably have to either use a tool like a shield or something that has like some of the stats and be able to basically CSV the data in. ⁓ or even just go manually, put it in for now. ⁓ but yeah. Jay Nathan: Well, so if you do with cloud code, it can actually script all that out for you and do everything, right? So it essentially, it would write a Python script, which would automate the browser and go call up the page, make it probably do the whole thing, right? I bet you could get it to the whole thing, but that's sort of the fun of this is like what you're doing first, this is MVP, right? This is like the minimum viable thing to... Jeff Breunsbach: find the posts. Yeah. All right, that's my. Good. Mm-hmm. Jay Nathan: the minimum viable valuable thing for me personally. And that's where this explosion of creativity is coming from, right? Because you would have never sat down to do this two years, three years ago, right? But now you've got the seed of a really cool product idea, right? And you start to see, well, okay, well, this is great, but it's all manual. Now I need to go connect it to Claude. And now I need to go, you know, automatically post into LinkedIn and then automatically come get the stats from that post and bring it back and. And I think, you know, that's where these things, they blossom over time from a, a, an iterative motion, as opposed to like having to dream up some big thing and then go build it. And then you get it back. And it's like, that's old school product development. This is new school product development, right? Which is ⁓ dream it up ⁓ then I start iterating on it. Right ⁓ ⁓ ⁓ real products, that's dangerous. ⁓ the reason ⁓ because. Jeff Breunsbach: Yeah. Jay Nathan: If you're shipping a real, quote unquote real product, it's gotta be secure, it's gotta be ⁓ multi-tenant, it's gotta be bulletproof in a number of ways, right? If I'm selling this product, and that's where I think there's a difference between just pure ViDECoding and then using Claude code to engineer software. But anyway, we can go down that path some other time. Jeff Breunsbach: Yeah. Yeah, no, makes sense. Okay, so I've already got homework, which is I think I'm probably gonna have to put this in cloud code. And this will be my first experience. Like, I've mainly done all my stuff in cloud co work. And it's been like, pretty good, right? I've kind of gotten everything where I need to in terms of like some of my day to day work. ⁓ But I think it'll be fun to go mess around cloud code, specifically. And I got two ideas for you that kind of came out of this, which I'm curious to see. So you just built your whole like, pipeline view. ⁓ how you wanted it and what you needed. So as I kind of took that idea and then this, this kind of day to day flow. And what I started to think about was like, actually could I develop a, ⁓ simple application for my CS team about like what they're supposed to be doing on certain days of the week? Do they have meetings? Do they have, basically can you create a CS cockpit that pulls in data from where we need it to, but it's basically, but it's basically more specific to like what we need, the views that we need, the types of things that they need to be able to see. Um, you know, instead of, um, it gives them a little bit more interactivity of like certain tasks or things that are up here already. Like, so is there a way that you could, you know, essentially like, could this become a hub for like your CSM team? Um, where you kind of put, Hey, here's not only here's customer meetings you have, but here's tasks that you have to get done this week about certain things or, we got to go share this type of update with customers. But this becomes like similar to what you've done with HubSpot. Could you do the same thing with like, uh, instead of feeling locked into somebody else's architecture, you kind of create your own. ⁓ day-to-day planner, day-to-day workflow, like meetings and notes and all this other stuff that you feel like is important. And then it all still gets logged back into the right systems and tools that you use. Jay Nathan: Yeah, I mean, that's it, right? And I think I'm starting to think of agents as in, I'll give credit to Jason Lemkin on this one, but like an agent is almost like an application designed for a job. So my customer success agent is ⁓ doing things on my behalf and pulling things together. I don't know, I guess why I'm hemming and hawing a little bit here is... I think that, you know, there's some of the stuff you wanna build for yourself, but some of it, to your point, is a team thing, right? And ⁓ because like what I built for the pipeline, stuff is great, and it's just like, it's filling in a gap, right? There's little cracks between all the software that we use. This is filling in a gap. The core data is still in our CRM system, no matter what. It will always be there. But. some of the user experiences that sit on top and the specific way I want to use all the data in that core reliable system, those are nuanced. ⁓ I don't think they should be limited to just me, right? They should be sort of developed in a way that's a little bit more deliberate and that's designed for the people in my company who do that function, right? So you get back to having custom software, right? That's what this ultimately is. ⁓ Jeff Breunsbach: Yeah. Jay Nathan: And, but, but the question is like, I don't know, I'm getting off topic of your, of your point, I think, but like you, do you really need to go buy packages, software for everything now? Right? You don't. Jeff Breunsbach: Yeah, that's a good question. Yeah. ⁓ and I think like this to your point though, I think what's, what's opening up though is that you start to see, ⁓ I guess you're starting to understand, or at least I'm starting to see and understand like, okay, where does it make sense for us to essentially try and like you said, fill in the gap with AI and like, what are we really trying to solve for in the short term in the immediacy and like, it help us do that? ⁓ and like how big or how, how much time would it take us to get to something that's like V1, V2, V3, right? Like how long would it take us? So, you know, like an example for you is our CEO basically built a, a ⁓ whole environment for onboarding ⁓ where we now can basically walk somebody through onboarding. And it, and it actually is like, as if we were in their application and they were calling our API. And so you can actually see what happens in your application. So I have a small little like window that shows, okay, I'm going through a patient workflow, right? So, Hey, I'm checking out to buy something. And then on the right hand side, it's showing, okay, if you were to call our API, here's basically the metadata that's coming in and out of those calls. And so you can see basically like, okay, you're pressing this button and it's calling junctions API. And here's what is it, here's what it is going to return for you. And it's like a live, it's a live environment so that they can see, okay, like Jay Nathan: ⁓ that's cool. Jeff Breunsbach: Where do I need to architect some of this? like for onboarding, it's become, uh, it's something that I think we're going to start to use because it's like, okay, I can actually show them what they're supposed to be doing and where they're supposed to be doing it. Like in a specific way. Um, so it's like another, I don't know, example of just like something that we're w we've been working on, um, with like, do we, how do we infuse AI and some of these workflows that we have? Um, and like this one specifically is around onboarding. The next step for that would be like we have to capture some information from the customer. So can we actually just build that into that experience? Can we just capture that information in that? Because it's a link that we can send them and can we basically specify that by customer and then be able to capture certain information that like tracks back to us. So anyways, this is just another example of like something that we've been working on to. Jay Nathan: It strikes me that... You know, you always buy software for the common things that need to work in a similar way across companies. The reason it's been difficult, so I'll just say this plainly, the customer success platform category as a whole is stalling. Like the biggest player in the market has not grown in multiple years, and you can go look that up if you don't know who I'm talking about. ⁓ I think part of the challenge is exactly what you just said. Every product is different, the market it serves is different. So there's this nuance layer when you talk about value realization for customers, when you talk about onboarding for customers. The layer of things that are common across all software companies is very thin relative to the domain specific knowledge, the product specific. workflows that go into actually deploying it. So that's why I think, you know, the world migrates back toward just core CRM. And then what you're describing almost becomes like part of the product, right? And all of a sudden onboarding can be much more streamlined and smooth because the product team is building that as an experience. It's part of the overall. Jeff Breunsbach: Yeah. Yeah. Jay Nathan: way that you set the product up, not some third arm off to the side that is being controlled outside of the product or not fully integrated with the product. But it's bespoke, and that's the secret sauce, right? Jeff Breunsbach: Yeah. Yeah. And I like to even to that point, like I'm actually meeting with our product mark, product manager and an engineering lead to, cause we've all been circling around how do we improve onboarding. And it's not to your point, it's not just a CS can solve this problem, right? It's like, how do we actually improve ⁓ the product side and the, like, essentially the flow of that onboarding in a way that like, we could actually, could we actually embed that entire experience into the product so that that's not even a separate URL. It's not something we walk them through. It's little like, Hey, why don't you log into your account, you pull it up and we'll walk through onboarding together for a few, like I'm going to talk you through this and show you a couple of things, but like it'll, gives you some of the optionality, I think. And to me like that is those, like, guess getting the customer closer to the real thing. Closest to the real thing is like how you're going to reduce time to value. Jay Nathan: So yes, that is super powerful. Problems you never would have actually walked up to previously you can now walk up to, right? Because it takes so much less time to generate code. Now you do have to maintain the code, right? And that's real. So don't wanna forget that. But then the question is, well, what does the team structure actually look like? Jeff Breunsbach: Yeah. Jay Nathan: for doing this because what you want is exactly what you just said, right? You want domain experts closer to the product development because guess what? Every engineer that's tried to design an onboarding flow from afar, I won't say every, but if you try to design an onboarding flow from afar, you're probably gonna fail because you don't actually understand the customer. You don't actually understand where their data lives. You don't actually understand who they have to talk to to get, you know. Jeff Breunsbach: Yeah. Jay Nathan: all the different pieces done, like you just don't understand, you're not close enough to it, but the people in customer success are. that is a really powerful way to like, like everybody's in product development now because you got, and Jeff, you and I have talked about this for years, right? All the different customer success processes and flows like onboarding, you should think of that, like your company, those processes and flows are part of the. total product, the full product, right? And before they were human driven, now we can actually make them a little bit more product, like actual product driven, even in a high touch scenario, we can build this into the product. Cause the product is not the software, it's the outcome in the way you get there, right? That people, that's what people are buying from us even more today than ever. Jeff Breunsbach: Yeah. It's, I forget if I mentioned this last week or not. I don't think I did it, but, I think this is the first, this is the first time in my career that I've actually built a customer journey that is externally focused because what I've, what we're doing right now is we are essentially our customers, their journey is actually at the end of the day, their patient journey, because they're essentially trying to build a product that's. so what we've done is we've now basically said. Okay, we're going to show you where you use Junction as a product in your patient journey, which to me is the ultimate like, okay, because that's what they care about most. And that's ultimately what we care about most. And so now we've been able to basically say, and just that experience I described, right, is like, okay, we're I'm showing you an example or sample of a typical flow that you might have for one of your patients. And I'm showing you exactly where Junction fits into that flow for you. And then like, I can visualize that for you. And it's like, to me, it's one of the first times where Jay Nathan: Right. Jeff Breunsbach: I've actually felt like, we didn't just build a customer journey that says like, okay, I need you to fill out these seven fields. And then I need you to tell me this so I can get you a provision and account, right? Like none of that exists in this kind of new flow that we're looking at right now. And it's to me, pretty powerful. If you go to a customer, like, just like you said, right, it's pretty cool where you're like, like you're showing me how to use your product, but it's in the basis of like what I'm thinking about most. Jay Nathan: I call, I have two different words or phrases for you've heard this before from me, but one of those is a service blueprint, like what you need to do to interact with my company. And the other is customer journey. Like how do I mature into utilizing, fully adopting your product and getting better at what it is I'm trying to get better at by using it. Service blueprint versus customer journey. And I think if you, like where those get confused is where you just. I think that's the difference between like truly pursuing customer success as an outcome versus running an operational team of CSMs to do. Like if I'm a CEO, I want to hear much more about the customer journey, not the service blueprint. I assume you can figure the service blueprint out, but do we all agree? Cause the customer journey is actually part of the go-to-market strategy, right? It's like, here's the problem, presale, they're aware of the problem. Jeff Breunsbach: Yeah. Jay Nathan: We make them aware of our solution. We deliver the solution and deploy it. Then they get value out of it. Then they become an advocate for us to teach other people how to solve that problem. That's the customer success flow. And that is the customer journey that our customers are on. Jeff Breunsbach: Yeah. Yeah. all right. ⁓ maybe I think we've got a couple of minutes left. Can I tell you about another example that my team is working on right now, which I thought was pretty cool. Cause we talked about team, I think two or three episodes, we kind of talked about like team wide AI adoption. ⁓ and one of the things I've mentioned is that like we're spending one day a month right now, at least, ⁓ where we basically you get to, you get your team together at like eight or nine in the morning and you basically say, okay, what are we working on today? And we have to basically automate something or Jay Nathan: Sure. Yeah. Jeff Breunsbach: use AI or automate something by the end of the day that is like something that is in our department. That's like something that is a pain point for us that we feel like can be solved by software. And I let, I would say like up until now I've been, I've been the one who's driven a lot of like the AI stuff for our, our team. And so I purposefully kind of said, okay, you guys decide what you want to do. Like I want you guys to start to be more AI curious and you guys figure it out. So what they came back with was this idea of, ⁓ Right now we've just you know, we're tracking renewals and what they want to do is they want to get a better pulse on a weekly basis of what's happening with their accounts ⁓ that are specifically up for renewals that they're working on. So what they're doing and what we're working on right now is we've hooked up the MCP for plan hat ⁓ plan hat right now, then we've got access to a bunch of our tools. And so ⁓ what they're working on is essentially a way that it will go through support tickets, emails, Slack messages, ⁓ fathom call recordings, and then it will summarize those things, and then it will drop it into a field inside of PlanHat that's attached to their opportunity record for the renewal. ⁓ And so it'll start to do that on a weekly basis. So every week they'll be getting a pulse of like, okay, what's happened over the past week with this customer? You know, did I have a call? What did we talk about? Was there support tickets coming in? Did they have any slack threads that were like, you know, important or something that I should follow up on? So now they're essentially going to get a digest every single week on their opportunity record for renewal for their customer. That is just AI driven or AI written and, ⁓ but summarizes, I guess, like all the aspects that we touch a customer through right now ⁓ across the board. Jay Nathan: I love it, man. Jeff Breunsbach: I thought it was pretty good. thought for like their like first time, I thought it was like a pretty good example of like something that you could go build like, you know, and they're like, Hey, this will give me context so that I can, because essentially what we've started to do right is like, what's your key next step on that deal every single week, we got to close it, you know, we're trying to close renewals, like four weeks in advance of the actual renewal. like, what are we doing? And so now they've got this context built in where they can then And so maybe we automate what the key next step is, you know, the future or, you know, use AI to help us. But at least they now have the context like fully built into somewhere that they can look at and then they can write what their next step is as they like plan out their week. So I thought it was pretty good example so far. Jay Nathan: Yeah, absolutely. took you a day to build that and it's in production for all accounts and all opportunity renewal opportunities. Now, you know, you just you just described staircase AI, right? Like that's all it does. I'm not to belittle it, but ⁓ that's that's what it is. So ⁓ I think that's really cool. That's amazing that they were able to do that. Now, did what IT dependencies did you all did your team have on that? Jeff Breunsbach: What's that? Yeah. Yeah. So we're still working. we've got it. So we can actually hook that up in a workflow inside of plan hat. So plan hat has a workflow feature. They have AI nodes that you can build in. ⁓ So the thing that the two things that are dependencies right now is that we have to get our slack messages into plan hat. ⁓ And so I'm working on a way for us to do that. It doesn't seem very hard. I just have to figure out the rights because I don't want all slack messages because we have slack. We have slack. ⁓ Jay Nathan: Ahem. ⁓ cool. Okay. channels per customer. Okay. Jeff Breunsbach: channels with all of our customers. And so, ⁓ but I'm trying to work on a way that, know, essentially, hey, if we think this thread is important enough, push this into plan hat as basically a context. Like I want this context in the account record is like how I'm trying to get my team to think about it. ⁓ So there's that dependency and then pylon tickets, we use pylon for our support. So those are two things pylon already has ⁓ a way that I can like easily set that up. So that's, I don't see a big issue there. So those are two things, but otherwise, Like we've already got email tracking. ⁓ I've already got ⁓ invoices for ⁓ AR, accounts receivable. I've already got invoices going through. I've got, ⁓ what's the other one? Fathom call recordings ⁓ and like meetings on the calendar ⁓ are all kind of baked in. We're starting to do project plans. ⁓ Jay Nathan: And all that stuff gets centralized into PlanHat. Jeff Breunsbach: Yep, right now. Jay Nathan: So let me ask you this. Why doesn't it get centralized into Salesforce instead? Or what do you guys use HubSpot or Salesforce? Jeff Breunsbach: We use HubSpot mainly because I, I would say mainly because the, like our team, I think the UI functionality of HubSpot was like a detractor for us to actually using the product. And I don't think our team was going to like fully go into like building applications on top of HubSpot or something quite yet. So I went to look for something that integrated with a bunch of the tools that we had. And I would say the second thing that was like a factor for me is that ⁓ one thing I do like about plan hat is you can have environments for your customers. ⁓ So we can actually spin up. ⁓ I can spin up a Balboa environment that has ⁓ your onboarding deck. It's got your any sort of documents I can, I can. Yeah. Shared workspace. And so that includes, includes though, like one of the things that I keep thinking about a lot is like, ⁓ that includes. ⁓ Jay Nathan: Feels like a workspace, like a shared workspace kind of thing. Jeff Breunsbach: project management tracking back and forth. And so I can actually tag you in something you can tag me in something. ⁓ and so it's a, and then that actually gets back into plan hat and into a central place where they can then see all their tasks and everything. ⁓ a couple of those things were to me important for us to try and get the team to use something centrally. ⁓ and I don't think I would have been able to do any, like a lot of those things inside of HubSpot. Or if I could, I think to your point, I would have had to buy up into different, like we use HubSpot mainly for Jay Nathan: Yep. Okay. Yeah, yeah, yeah. Jeff Breunsbach: I was CRM right now. don't, you know, we don't have service hub. We don't have success hub. I don't have like a bunch of these other things. And so I felt like I would have had to basically go spend more money with HubSpot in order to do this. Jay Nathan: Yep. That's all that's all true. ⁓ Okay, cool. Sorry, I totally put you on the spot there. Jeff Breunsbach: All right. No, ⁓ I've had to be put on the spot lately. You know, I mean, I will say too, right? Like I think the other thing that I've just continued to realize is that like, this stuff all takes time and like you've got to prioritize it. Like I've, I've, ⁓ I've come to realize that I'm just not prioritizing it enough. So it's been a big, like right now over the next six to seven weeks, it's like a huge sprint or huge couple of sprints for me of like, what are we actually getting into my team's hands? That's going to help them do their jobs. And like, it's all got to be baked back into ⁓ plan hat. Jay Nathan: Yeah. Yeah. Jeff Breunsbach: So. Jay Nathan: Well, plan hat to me sounds like where I think probably all products have to go is you have to have the ability to build agents in every software platform because again, it's back to the custom pieces that make you unique and special and that make you unique to your customers because all the common stuff is just service blueprint shit, right? The customer journey stuff is all the best bespoke stuff and that's Jeff Breunsbach: Yes. Jay Nathan: that's the stuff that really has to be designed uniquely for you. so agents basically give you the ability to create those unique kind of integration points with the customer versus just having another work management tool. Jeff Breunsbach: Yeah, I know this is probably a very simple use case that's like been solved years ago, but I'm working on it right now, which I thought was pretty fun too, is I am working inside a plan hat to effectively go through all that same context, and then be able for somebody to ping it in Slack. And so like, I want to know it's half, hey, what's the most thing that's happening in this account? And so it'll put like, to your point, it's taking in all this context, and then, ⁓ and then I'm having Claude write up a summary based off that. So it's not just like, here's the meeting dates and here's what, know, whatever. It's like an actual summary that says like, okay, we've had this many meetings and here's what we've talked about and here's who's attended those. And it's like more in a style. anyways, like I think getting more and more into this like true workflow CS stuff has been pretty fun the last couple of Yeah. Jay Nathan: It's awesome, man. It sounds like you guys are doing some great stuff. ⁓ It's gonna be... Yeah, I mean, that's it, right? Taking it out of ideation mode and putting it into production is the key for all of this. like we talked about before, getting it out of my laptop and into a place where the entire team can use it, it's key at this point. So. Jeff Breunsbach: Yeah, just got to execute on it. Yep. Cool. All right. Well, that was some fun stuff to show you this week. I don't know if there's anything else you want to cover or if Jay Nathan: Yeah, man, need to, no, need to, ⁓ I need to get on my A game. I was out for a few days last week and ⁓ as you know, and ⁓ I actually had a nice little break from the rat race of AI for four days. So my brain's starting to catch on fire again now thanks to this call. So it's good. Jeff Breunsbach: Yeah. Cool. All right. Well, we'll see you on the next episode. Appreciate y'all for listening. Jay Nathan: All right. See you.