Jeff Breunsbach: Welcome back to the Chief Customer Officer Podcast. Jay, welcome back. How's it going?

Jay Nathan: Thank you. it's going great. How are you?

Jeff Breunsbach: It's I like to think of it as a chaos factory right now, but you just keep one foot in front of the other. Yeah.

Jay Nathan: Yeah, control chaos. That's right. That's right.

Jeff Breunsbach: But we I had an off site with my CS team this week. And so I thought it would be fun to talk about at least one thing, one exercise that we did that I found pretty enlightening. And I I I feel like it's probably extremely basic. But I thought you would enjoy like talking through that and maybe like what we learned through that process and trying to help my team understand about like

Jay Nathan: Yeah.

Jeff Breunsbach: why we did it.

Jay Nathan: Absolutely. I you know what? The basics are good. I

Jeff Breunsbach: Yes.

Jay Nathan: think I think I think a lot of people make this stuff way more complex than it has to be. So let's hear it. I want to hear the basics.

Jeff Breunsbach: Yeah. so I forget exactly what I called it. I'll have to look up what I called the exercise. I thought I had a name for it. And but what I had my team do was literally I sat at the front of the room and I had a little notion doc up on the screen. and I had them just start list I had them go to their calendars last week and I just said start listing things that you did with customers, meetings. Slack messages, emails, but like, what was the thing that you were actually doing? And so what started to come out of that was you know, maybe we were do we did a roadmap discussion with a customer almost in reverse. We actually heard we they were telling us their roadmap, like what products they were launching this year. That's a cool thing that we're doing. We should probably do more of those. I had people tell us about, we configured somebody's team settings or team name or org name. And it's like, well, wait a minute, why are we, why are we doing that? That seems So it's super interesting because we started to just get a laundry list of things. And then we wrote those things down on sticky notes and you know, just kind of made it a little bit of a hokey exercise. But like we listed out five columns on the table and it was keep it, start it, keep it start it, automate it, delegate it, cut it. And every sticky note went into a certain area. And the started bucket was, you know, we started thinking about things that we weren't doing today that we just know we wanted to. And so we started adding some sticky notes there. But the again, I like I mentioned, I don't think this was a, you know, particularly novel exercise or something that was super complex. But it was just, I think, almost like a from a change management perspective, I thought it was enlightening for our team to go through because if you looked at the sticky notes, if I was to put up the five columns, right, the things that were in keep it that we do today was like. Very minimal, which I think again of like the CS teams, you know, like you try to almost you almost I think instinctively try to defend everything that you do because you're like, okay, if you take something from me, then what am I doing? Right? Like, my God, like what value am I providing? And you feel almost like kind of like naked, right? You feel like, okay, you've taken something from me and and now I'm more vulnerable inside of the company because like what is CS really doing with the customer? and so I think like your natural place is to put a lot of stuff under keep it. but I think we had I think we probably had twenty five or thirty sticky notes, maybe more, but I mean, fifty to sixty percent of them were either under delegate it or automate it or cut it, you know.

Jay Nathan: What what is what's the number number one objective of your team from a business results standpoint? What are they responsible for?

Jeff Breunsbach: Revenue.

Jay Nathan: Okay, and when you say revenue, break that down a little bit more.

Jeff Breunsbach: renewals and then upsell, cross sell. I mean, we're early stage company, so we need to be contributing to the top line revenue of the business and making sure that we can show, you know, we've got stability in the market, we can grow customers over time. Like that story is important for us right now.

Jay Nathan: So I I'm sure like that is the answer your CEO loves as well, right? And your your board and everybody because that means everybody is focused on growing the business, right? And so that is a nice anchor for what you did too. It's like it's not just opinions, right? It's it's like if we want to do more of this, and I think if I if you I won't I won't I don't wanna I won't share anything you've shared with me privately, but Seems like your C S team is driving as much or close to as much revenue as your sales team is in terms of expansion. Right? Like if you want that to continue, which everybody wants that to continue, that's a net revenue retention, then here are the things that we probably should figure out how to delegate or automate. I love those categories, by the way, versus doing ourselves. So

Jeff Breunsbach: Yeah. I I think that was and I think that was like the change management piece for the team, right? Is that we we have been growing over the last two quarters in in the upsell piece and the renewals, we've been taking on more revenue responsibility, almost while being weighed down by these things that are in those like automated and delegated buckets. And you're like, okay, like imagine if we're able to essentially release ourselves of these, right? Then then We kind of get more opportunity to go find those opportunities, probably. I think we actually discover more. we get more time to be strategic about those, right? We actually get to spend more time on those deals themselves. Right. Like right now, it actually feels rushed. We're getting them done and we're doing things in in a in a way that is constructive. But again, I would say like we don't have like ample time right now to sit around and really strategize. Okay, is this the right deal to put in front of that customer? It's really like a Bang bang go. And you know, like you would actually probably could have more time to think about these things and and like think about timing, right? And and I think some of the nuances that you want your team spending time on, even though people don't think about those, I think like those are some of the things that you would probably get time back on if we're able to move those things off, you know, delegate it, automate it, cut it. But I think the other thing that's interesting is that, like you said, even with this goal in mind of hey, we we want to be really revenue focused right now, it's important for us. You still end up like under the start it bucket, right? You start recognizing that they're like again, this is all a leading to lagging game, right? Like the lagging

Jay Nathan: Mm-hmm.

Jeff Breunsbach: game here is the revenue piece. So I can't say, you know, so like when you're thinking about okay, like what activities am I gonna go do or what what do want to start doing, right? It's not like find more revenue. I don't know, it's not like renewals is

Jay Nathan: Yeah.

Jeff Breunsbach: like the start it, right? No, I need to be getting upstream of that. So like another example is like this. we had a customer come to us and Showcase their roadmap to us. And the way that we position that was, hey, you're on a great, you're on a great trajectory right now. You're growing. We could help you under like we could help you understand where our product fits in even more if you allow us some latitude to understand what's coming next quarter, what's coming the c the quarter after that. Right now we're reacting to you and we'd rather be proactive with you. And you should have seen how they were like, wow, that's a yeah, you're you're right. We should be like collaborating on this more. And it just opened up a a great conversation. And now we actually have. A shared document with that customer with their initiatives and our initiatives basically of the products they're going to launch over the next two quarters so that we can understand where they are. And then like we've shared that with our product team to say, hey, are there things on this list that we don't support today that we can be getting in front of you and getting ahead of? And so it's just created a much different conversation. But to me, that isn't a hey, it's gonna drive revenue right now today, right? But that's the lagging of like, hey, if we help them grow, if we help get these things in motion, then That's going to turn into revenue, which is a great thing for us and it's gr it's all kind of natural growth.

Jay Nathan: And that that's a really good leading indicator. Like huge. So if you if you do that two or three times with two or three different customers and you think it actually works, then that is a way better s quote unquote playbook item than a QBR, right? It's like, no,

Jeff Breunsbach: Yeah. Yeah.

Jay Nathan: we have something called a reverse roadmap review, right? And

Jeff Breunsbach: Yeah.

Jay Nathan: we're gonna measure how many of our top 200 customers, we've done a reverse roadmap review. And that is a leading indicator. Because if we do that reverse roadmap review with our customers, they are X times more likely, we believe, to that's the strategy, right? Which is we believe if we do this action, it's going to create this kind of result in this kind of time frame. That is really, really powerful. I love that. And that's a specif specificity. This is why customer success isn't one thing for every company.

Jeff Breunsbach: Yeah.

Jay Nathan: Because those little nuanced motions are what separates your company from the next company that does what or tries to do what you do. Right. And by the way, that some of that might be your secret sauce too. You don't want to talk about that publicly. So yeah, if you're a competitor of junction, strike the record on this one. But like that's really cool. I I love that. And and that is a great leading indicator.

Jeff Breunsbach: Yeah. So it was f I mean, I it was a great I I think at the end of the day, like it was again, it's it's a simple exercise and I think sometimes it can be overlooked. It it almost reminds me of you know, I think there's some almost like leadership coaches out there that still subscribe to the like, you know what, go to your go to your week last week and and you know, categorize all the time you spent, the meetings you had and tell me where you spent your time. And I still think that that's like leading leadership coaches I think still use that exercise as like You told this is the most important thing. Where are you spending your time? Right. And I think it's like a similar, it's similar exercise for the CS team and

Jay Nathan: Yeah.

Jeff Breunsbach: something that you should just continue to do over time. And the thing we talked about, I talked about with my team too, is that once you take something on, it's really hard to take it to take it off, right? It's just a natural habit of okay, I'm I'm doing this, I'm comfortable with it. Okay, well, I don't want to go, you know, I don't want to give it to that person. I have to go train them, I have to go do it, or I don't want to cut it because then what happens if I cut it and somebody needs it, right? Like they're So, like naturally, our instinct should be to not try to do something recurring. another example,

Jay Nathan: Right.

Jeff Breunsbach: right, right, right now we have we have some customers that we do check-in calls with. And so one of the one of the goals that we talked about as a CS team at this offsite was could we cut or eliminate 50% of our check-in calls right now? Like in this quarter, can we just go to our customers?

Jay Nathan: And those are like what, Jeff, weekly or bi weekly or monthly calls with the customer to quote unquote check in. Yeah.

Jeff Breunsbach: Yeah. And yeah, and it's and it's happened naturally, right? Because we have onboarding and it's basically just been a a continuation of onboarding. Hey, this call feels comfortable for both sides. The customer feels like, okay, I have a connection point with you. We can talk about things. But literally we have fathom for all of our calls. And so, you know, what did I do? I went and and I just checked over the last couple of weeks and just I watched a handful and it's what you'd expect. We're reviewing support tickets. They're asking questions about the product. It's like product enablement and like It's probably support documentation that we should or product documentation that we should have. And you start looking at it and it's like, okay, we just spent a half hour. And going back to our our kind of point about, you know, some of these exercise or some of these more strategic things, right? We just spent a half hour with the customer, didn't really further the relationship, didn't really talk about anything strategic. We just made them,

Jay Nathan: Yeah.

Jeff Breunsbach: you know, we just kind of listened and made them feel better, which again isn't it isn't wrong necessarily, but it's just not gonna get us far. Like it it just

Jay Nathan: Some valuable.

Jeff Breunsbach: Yeah, and it's it's a it's a almost like a short-term fix for the customer, right? Like, okay, they feel better. They can go talk to their boss and say, okay, I talked to, you know, I talked to this person at junction, you know, everything's okay. Cool. so

Jay Nathan: Yeah.

Jeff Breunsbach: yeah, so I so we just set out, you know, can we go eliminate 50% of our check-ins calls? and and to me, like that isn't just going to just say to the customer, hey, we're eliminating this, right? But no, go back to them and say, Hey, we don't have an an active launch going on. We don't have you know, a a major project that we're like leading towards. you know, I think right now, this time it was probably better spent for me doing X, Y, and Z on your account, or, you know, for you to be spending time doing X, Y, and Z. There's there's a way that we can go position this in the right way so that it doesn't feel, you know, the customer doesn't feel slided by it and they they feel like they're getting less service. But yeah, we should we should try and figure out what is that right positioning and messaging. And we should go try and and drive that to make it to make it a part of our like Playbook. Hey, after implementation, there's a reason we have a call. And it's because we've got like a project ongoing. We have some sort of, you know, finite time that we're trying to get to. We're leading them on something. Hey, they're implementing something of our products. Like there's a reason, but like we should set a start and an end date to that. And if we're at the end date and it's still not done, okay, we can continue it. But if it's at the end date and it's done, cool. We're gonna move on and

Jay Nathan: Yeah. And there's a the the way to sort of facilitate that is with a transition call. Like, hey, this is going back to your team now. And by the way, like to your point, it's not just, hey, we're not doing this anymore. You might need processes you don't have today, meaning the customer should be able to check self serve check on their own support cases via portal. Like, come on, man, right? Like

Jeff Breunsbach: Yeah. Yeah.

Jay Nathan: it's 2026. Is that not possible? And then to your point, like the example of the reverse product roadmap review. Like what is the specific thing? If we're getting on a call together, it's not just a check-in. There's a purpose for it. Right. And it's like, hey, we're gonna accomplish this thing. In my world, customers pay for that. The that high touch one-to-one kind of model. not not to say you want to get into the services game, but that's that's a that's a paid high touch service, right? Everybody else, the default is you use our portal. You know, we do specific things that are strategic to both of us, like to advance the relationship. I think all this comes back to like you mentioned consultants doing time and motion study kind of work. It all comes back to that's fine, but you've got to do that kind of work with a lens, like a go-to-market lens on it, which is why it's so important that your team is tied to revenue, and so important that that's the anchor, which is what are the activities that lead us to. Growing the relationship from a a contractual and revenue standpoint. I don't know if you remember this, but we worked with a back when we were consulting with Customer Imperative, you and I worked with a with a company. I don't know if we worked on this one together, but they had a it was a very high-end product. They sold it to, they had like 200 targets globally. They knew exactly who they were. It was a very specific set of accounts, both prospects and their existing customers. Like it was not. Ever destined to be a massive company, but every deal was multi-million dollar deal. And they had a really good marketing guy who wasn't a marketer by trade, but he just had very good intuition. And he created this whole program of how to go indoctrinate those 200, those 200 companies that were their

Jeff Breunsbach: Yeah.

Jay Nathan: targets with the enablement that they needed not just to use the product, but to actually buy the product. In the first place to understand the problem. This were for big contact centers. And they solved the problems, some problems for those contact centers in a very specific and scalable way. It was a very cool product, very

Jeff Breunsbach: Interesting.

Jay Nathan: direct ROI. And he created this account-based marketing program where they went out into the field pre-sale and brought customers into these workshops where they created where they basically they he called it giving them the flu. This was pre-pandemic, so it was a little more kosher back then. But flu stood for field level understanding. And they tracked for every single one of those accounts how many people within each account had the flu and had an understanding.

Jeff Breunsbach: Interesting.

Jay Nathan: And then they would run workshops to have them sort of map this stuff out. And then they would use those use cases that those executives mapped out for their business to then go sell them the product. And then they so like.

Jeff Breunsbach: Co

Jay Nathan: That's, you know, that's the kind of stuff that is separates a real go-to-market leader from somebody who's just, you know, a delivery leader keeping the wheels on the bus, right? Like that's actually, you know, creating value for the company, creating some depth and contour to your engagement model with a customer that, you know, just, you know, doing touch points, having check-ins, measuring, you know, generalized activity doesn't touch that. Right. It's not even close. So anyway, that came to mind

Jeff Breunsbach: It it

Jay Nathan: as you were talking.

Jeff Breunsbach: yeah, well it goes it g I think it goes back to your your earlier point as well, right? Of of each of these businesses are are again kind of unique creatures, right? And you've gotta figure like you said, like he could have gone in and just said, Hey, yeah, let's go two hundred let's serve two hundred ads on l Linked into these people and let's, you know, send them a gift or whatever, right? But he's like, No, no, we gotta like we have to go engage with these two hundred in a way that is like Like how are we going to do it as a business, right? Like what do we want to represent

Jay Nathan: Yeah.

Jeff Breunsbach: in the market? It's almost like a brand play where it's like, how do we want them talking about us? And like, okay, we're gonna go infiltrate and that's that's what is like gonna be the remembering part about what they're doing.

Jay Nathan: Yeah. I was reading about Palantir's agent camp model this morning. Have you seen this?

Jeff Breunsbach: it's interesting. I haven't seen it, but it weren't are they the ones who everyone I think is like coining that they kind of came up with the FDE model? Like are they yeah.

Jay Nathan: Yeah, they did. Yeah. They they were the first to well, I mean, for yeah, forward deployed engineer was a Palantir thing and they they've been doing that for, you know, ten or

Jeff Breunsbach: Got it. Yeah.

Jay Nathan: twelve years. and now everybody's jumping on that bandwagon. But they run something called agent camp. I haven't gotten too deep into it yet, but it but it's the same sort of thing, right? As the the field level understanding, the flu that that

Jeff Breunsbach: Yeah.

Jay Nathan: that our client was was, you know, give trying to give the market in in that. You actually had people mapping out their own use cases, right? And in it what it's doing at Palantir is it's resulting in much bigger deals right out of the gates. I think on their earnings call this week, they touted like a fifteen million dollar deal that started as is out of their agent camp. So I'll do some more research on it and bring it back. But like that's the kind of thing we're we're looking at, like we're doing with our clients now too, which is even even with you know Pendo being our primary partner, it's a very open platform, right? You can do a lot of different things. You can handle a lot of different use cases with with Pendo, just like you can with an agent platform. So, you know, having that mapped out in sort of a pre-sales workshop is a really powerful, powerful kind of kind of concept. So anyway, just interesting how it all comes together, you know, it is

Jeff Breunsbach: That's yeah, that's like

Jay Nathan: customer success.

Jeff Breunsbach: I feel like that's the number one thing that I keep I've again, I've I've probably tagged like six or seven articles even just this week about like every I it's kind of funny, right? I I feel like you see the shift in like what people put out in terms of content and like it kind of I don't know, it's like kind of a natural progression, but you read one article about something all of a sudden now I'm like, Okay, everyone's on the same bandwagon, but it's this whole idea of like, Cool, AI AI is awesome, but everyone has realized, okay, how am I going to get My enterprise team, like how are we adopting AI as a as a business, not as like an individual using Claude, pressing buttons? Like there's a difference

Jay Nathan: Yeah.

Jeff Breunsbach: in those, and I think everyone keeps mapping out that that is like the largest hurdle to go after. And that like that is where, you know, like that is where there's like a a major difference in just like doing AI myself versus like the multiplayer model.

Jay Nathan: Yep. Well, it's a good segue too, because I've been you know, we are actively sort of seeking out platform partners right now for agents, agent builder kind of models. Like you can go build agents in cloud or open AI, that's fine. To your point, a little bit more single player. They do have multiplayer mode, but then you're locked into their ecosystem. We've talked about that before. Then you've got The hyperscaler, you know, Microsoft, Google, AWS, you've got those platforms. You can go build agents there. That's fine. It's a little bit more, you know, further away from the business user when you do that. But then there's all the there are all these other agent platforms that are, you know, whether you're in sales and marketing, well, I should say sales, marketing, customer success, all all of the, you know, the vendors who were selling software into those. Domains are now selling agent platforms as either part of their software or that is their platform, right? It was an agent platform for X, for customer success, for sales, for marketing. So I know you've been using Planhat. They're I think they're in the category of had software in the customer success space, have added agent builder in that world. HubSpot's done the same thing, right? They have an agent builder now. You've got all the core data in there, but you can also build agents for sales, for marketing, for customer success in HubSpot. but we've been looking at some of the new pure play agent platforms that don't have any of the software legacy and they're just pure agent. And it's been really interesting to see like what a what an agent platform for customer success looks like from the ground up. But curious where what you're what you're seeing. with regard to to that since you've been sort of playing with the tech.

Jeff Breunsbach: Yeah. well I guess a question I think maybe like a question that comes to mind for me about this has been if you kind of broke down what an agent is, or I guess like where what is the I guess what is the real problem, I guess, when you think about agents that like these platforms are needing to solve. 'cause it feels like we've had kind of workflow builders before, right? It feels like we've had the ability

Jay Nathan: Mm.

Jeff Breunsbach: to kind of build out an automated process. It feels like okay, I can inject I can kind of inject AI into an automated process now. But I guess is the is the real, I guess, unlock when you think about okay, agents and these agent platforms is is it the ability to for I guess for context and accessing context and like making sure that that context is injected into those processes now? And then is it really I guess like the switching between models and finding efficiencies of of kind of models cost and like output. I I guess are those the two when you think about I guess agent platforms, is is that are those the two things that you feel like are I guess reasons that these platforms are going to go become popular?

Jay Nathan: It's really interesting because I feel like part of it, you know, we talk about a lot on this show. Implementing AI for AI's sake is really not is not the the goal anymore. It was maybe six months ago. Like everybody should be experimenting and we should all be doing something with AI. We still should be. But I I look at what some of the pure play agent platforms are trying to do. And it's like they're trying to r replace the whole thing with with agents when you still need some specific workflows, you still need some deterministic, you know, hey, we do this thing at this stage every single time. Like it's a blend. It's not just pure agent. And I think what's going to end up happening is the pendulum swings right to where, okay, we're we're going to use agents for everything. But at the end of the day, the pendulum's going to swing back, of course, like because customers are going to start saying, hey, look, like these agents are great, but they're burning a lot of tokens. And like, The outcome is a little bit different every time I run this agent for this particular task. So I just need that to be a specif. And then all of a sudden you have a feature, right? That is a defined workflow, maybe configurable, but it works the same way every time. It's code, right? Deterministic code, not an agent running. it may call agents for the for the part of that process that requires judgment,

Jeff Breunsbach: Yeah.

Jay Nathan: but it's not in and of itself, you know, an agent from end to end every single time. So I don't know if I answered your question there, but that's just like an observation.

Jeff Breunsbach: No, no, you did. Yeah, yeah. No, I think that's I think that's true. And I think that's where I'm getting to or maybe that's that's like what I'm seeing right now is that like you know, like I'm still building playbooks for my team to go execute on, but I want to go find the steps in those playbooks that I can automate that are I don't need an AI for it, but I know that like, okay, you're gonna go this is going to happen after these steps. And so I'd I'd rather like still I want you to spend as much time with our customer and to be thinking about like The right things. I don't want you to be inputting data or or changing a field or doing something on a record. So like where can I automate those? And so I think like you said, I I think there's some things that I'm recognizing. like right now, I'm kind of deep in what you built around call transcripts. And so I think there's super interesting things. Okay, like how can we take this call transcript that we again is is not about an opportunity record that we can then Can we judge it in a way that allows for the opportunity record to move to the next stage? Because the AI is determined, okay, look at all this, look at all these moments that okay, I can move it to the next stage confidently. and then, you know, that's on one end of the spectrum. The other end is like, okay, hey, after or hey, we need to go after this type of call that we have with the customer, I just need to go do this very specific thing and like it happens every single time. And like that, we could automate, doesn't need AI. But so what's interesting is like I'm actually feeling it's funny you mentioned like the pendulum swinging back because I feel like right now I'm in Planhat and one of the nice things is that I can do both of those things right now. I can inject AI

Jay Nathan: Mm-hmm.

Jeff Breunsbach: into specific steps of a workflow or I can just not even do that, right? I can just go through like and build an automation. so that's been beneficial from that stage.

Jay Nathan: Yeah, I mean, even even the the you know, progressing a deal to the next stage, like that's something you want to sort of be careful with because

Jeff Breunsbach: Yeah.

Jay Nathan: that affects your forecast. And a forecast is something that you as an individual need to stand behind. And if an agent just picks up on a signal from a call and says, this is not here, it's there, like, okay, somebody needs to validate that before I mean maybe you do do it automatically. That would I would argue that that's probably not the most high intensity. like workload that any sales rep or customer success rep has, right? But even, you know, to do it for them automatically introduces risk into your model as well. Non-deterministic risks. Like, okay, did it hallucinate when it moved it from discovery to proposal? Because I haven't actually provided a proposal yet. Right. So,

Jeff Breunsbach: Yeah. No, it's true.

Jay Nathan: and then, you know, when when somebody hits proposal or, you know, I put it into a different forecast category of, you know, Committed versus best case. Like that materially changes my forecast.

Jeff Breunsbach: Yeah.

Jay Nathan: So, you know, some things, you know, need to be structured. They don't need too much structure. They need just the right amount of structure. But those things, you know, inform all kinds of downstream processes.

Jeff Breunsbach: Yeah. What's interesting too about I think this again, I think we alluded to this maybe a couple of of episodes ago. but I I was reading an interesting article about kind of AI and and where it's going. And again, like this idea of the of like how it's going to fit into enterprise. And it was by Claire Vaux, and I'm like a total Claire Vaux fan right now because she has the podcast

Jay Nathan: Yeah, you are.

Jeff Breunsbach: How I AI. But I think one of the points that she was making was that one reason AI is kind of stalling at the enterprise level is because of the lack of creativity and the lack of like almost like product managers that are embedded on these teams that can creatively think or rethink current processes, think about new processes, think of you know, just thinking of solutions to, you know, workflows, problems, whatever. And so I I thought it was an interesting point because I like, again, I I think I'm playing with AI, interest interested with it, going down this path. But then to your point, right? Like I think there is this almost like block that is going to inevitably happen where you're like, okay, I'm almost too close. I'm into this job day to day. I'm too close to the problem that like I just know we do renewals. And so renewals happens in A, A, B, C, right? Instead of like almost being able to take the lens out and say, okay, I need a transaction which is a renewal, which is the end point. I don't know anything before that. So let's, you know, let's architect this almost like from scratch. So

Jay Nathan: Mm-hmm.

Jeff Breunsbach: I thought it was an interesting thing that she I think she put it out on Twitter basically about like what she's seeing from like a why is AI stalling at this I think kind of this single player multiplayer, like why is it really stalling of like, you know, getting into the enterprise more and really impacting maybe like kind of key workflows or anything is like the lack of creativity around, you know, in the idea of like kind of having an embedded PM, which I thought was a interesting point she was making.

Jay Nathan: Yeah, it's a it's a good time, in my opinion, if you are an actual product manager and not a program or project manager in product. I think it's a really good time to to be in that field because all IT is turning into product management if if you think about it, right? Buy versus build decisions, like

Jeff Breunsbach: Mm.

Jay Nathan: how do you create systems on you know, how do how do you create systems that That make the business better, that's a product problem, right? It's not a it's not a traditional IT problem per se. I mean, of course, you've got the traditional IT infrastructure security, privacy, scalability, all that kind of stuff, but that's secondary to like how are we gonna serve the business with all this spend? So, it's a really good point. So I know we're getting close on time, and and this is like a good probably segue to the last topic too. You've been playing around with this thing. called Buzz, which block put out Jack Dorsey's company. we're gonna use it. I don't know anything about it, so educate educate me.

Jeff Breunsbach: well I don't know if I'm the best to like about like educate about this, like because I I don't I'm probably not gonna use I'm not gonna use the right

Jay Nathan: Don't hedge, don't hedge.

Jeff Breunsbach: terminology, but I guess how I would coin it is it is a an AI or agentic messaging platform like built for agents to be using. And so you essentially so I guess would that be a harness? I guess it's a harness.

Jay Nathan: Yeah, it's an agent harness, for sure.

Jeff Breunsbach: So it's an open source agent harness that I'm self-hosting right now. So I'm I I bought us cloud space. and so we're hosting our own version of Buzz. And I guess the reason or what was intriguing to me and the reason why I went down this path, it was also kind of a bitch to set up, by the way. This is not like I was like,

Jay Nathan: yeah.

Jeff Breunsbach: I don't know, like it's you know, I've got a By storage space, I've got this is just stuff I haven't done. So I was curious. And I was like, hey, I want to go do this and learn and figure out. So part of this was just curiosity. But I guess the

Jay Nathan: Good, yeah.

Jeff Breunsbach: what intrigued me the most was the concept of basically every channel is an agent. And then you and I both can interact with those agents and like be molding them over time, be shaping them together versus like I have an MD file that you have to like check out. And like you have an MD5. I don't know. Like it just felt like when I first heard it, it just felt like, okay, I don't know. I've looked up things like, okay, do we have a shared GitHub repo that you and I both call for a shared skill? Or

Jay Nathan: Yep.

Jeff Breunsbach: do I go set up a Hermes agent or a whatever open claw agent? Or, you know, there's all these other things. And so I was just like, well, I don't know. You and I message back and forth all the time and like we're kind of used to Slack. And so like this just felt like a I guess a an easier experience for you and I both to just basically be like, Hey, so for instance, I think like I'm looking I I kind of have an idea of some agents that we want to spin up. So like the first one is like we do this podcast every week. Wouldn't it be great if we had kind of a podcast producer that's helping us, you know, frame what do we want to talk about? What are the topics? So we

Jay Nathan: We do, but it just stops running every time I close my laptop, so

Jeff Breunsbach: Sure. But but now yeah, so now it's

Jay Nathan: Mm-hmm.

Jeff Breunsbach: so now it's self hosted, so it'll run when our laptops aren't there. But now

Jay Nathan: Yep, it runs in the cloud. Yep.

Jeff Breunsbach: you and I and and your agent was like living on your computer and it was emailing us and I could email it, you know, and I was not engaging because also email is like tough for me. It felt too it felt too, I guess like not in live like not live enough. And so this is like a Slack experience where it's I we're messaging the agent back and forth and we can be telling it things. And again, you and I can kind of jump back and forth into it. So I guess like I've I've gotten it to a point where it's it's self-hosted. I've got an environment stood up. I'm starting to essentially just create channels for agents that we want. and that's kind of where I've I've I've gotten to is like that point. So I need to hopefully by the end of this week I'm going to have a handful of agents stood up. I'm thinking just two to three. and then I'll invite you to the space and then we can start shaping these things together. But basically right now, podcast producer, newsletter agent, and then actually one or I'm calling it a content I guess a content manager, but basically the content manager actually is going to orchestrate the podcast producer and the newsletter so that like our content is trying to be yeah, it's trying to be like how can

Jay Nathan: great. It all lines up. Yeah.

Jeff Breunsbach: we essentially start working all these things together? And so then the content manager will basically be doing that and facilitating with the other two agents because they can actually talk to each other as well. They can call the other agents from other channels and be doing this. those are like the three

Jay Nathan: Yeah. Of course. That's great.

Jeff Breunsbach: that that I think I'm gonna try and stand up this week and and get you added to.

Jay Nathan: That's great. Yeah. And so one of the things that I did have, so I was using nanoclaw locally to do something similar. But to your point, it's single that's still single player AI, right? It it's not multiplayer. And I'm sure I think you can set it up to be multiplayer, but that's beside the point. We did not do that. and I don't actually think it's tuned for that. I think Buzz actually is. But the cool thing when you really think about an agent. You want something that continuously learns. So the way this is a cool part of how I did set up that agent is that every time it produces a briefing for our podcast, it actually is consuming the transcripts of the prior podcast, even up to the one we did last week, and using that as input to the next thing so it can, you know, help us tie things together across the episodes. we haven't dialed it in yet, right? We had just gotten started with this, but we can do it now with the buzz agents. But Like that kind of thing is that's all work. It's it's sort of new work for us, right? Back to your point about how you design agents that take advantage of things that we would we would have only dreamed about before because we were like, Yeah, that's a great idea. We should at Of course you would never say we shouldn't go back to our old episodes and make sure we're tying new topics back to things we've talked about in the past. But who has time for that? Right. Well, now we don't have to spend the time on that. So, That's the idea. And and you could sort of extrapolate this back up to the earlier conversation we had too, which is those are the that's the same way your sales agents, your marketing agents, your customer success agents should be working, which is what are all the conversations that our team has had this week? How does that shape our messaging going forward? What are the key issues that have come up? What are the key learnings? What are the things we did that really landed well that we should turn into those playbooks that we talked about? Like that continuous learning loop is that's a whole new level of of, you know, iterative build on what we're trying to do, you know, in our teams as well. So

Jeff Breunsbach: Yeah.

Jay Nathan: I think it all, you know, all of what we're talking about sort of ties back to to what we talked about earlier with the with the customer success agents and even back to your team exercise that you did.

Jeff Breunsbach: Yeah, and one last thing I'll mention about that too, which I I I forget if we've talked about I think we've talked about this, but I have right now a simple MVP use case for the call transcripts that I've like taken from you is essentially I have a scheduled run that runs once a day and looks at the calls from yesterday and pushes it now into a channel that I'm just calling customer quotes. And so at first my MVP version was basically just like, hey, look for a customer that's calling us out in like a kind of a nice way it says, hey, we love junction, you know, you guys are doing great stuff, blah, blah, blah. Well, now I had it go train on our API docs and look at the different parts of our product. and then now so now I've got it like multi-category. And now it's like looking for quotes about each of our products and like things that like customers are saying good and bad and pushing that into the channel. And what I just thought was interesting, right? Is I I'm and I'm not really like tagging people yet or anything, but like I've started to notice there's some like engineers who are going in there now. And again, it just kind of reminds me like how removed engineers are typically from customers, right? Because they're kind of on the very back end

Jay Nathan: Yeah.

Jeff Breunsbach: of the process. They're building things. They're not hearing direct feedback or quotes. And and so like one of the things I just think about now is like, man, this is just a small thing. But like it's kind of cool just to watch the engineers now go. Now they're writing little comments on the threads, like, it's interesting that they mentioned this, this, and this, or like it's interesting that they said this. We actually can do that, you know, like you should go tell them we can do that. And so Again, it's a small trickle, but it's just like an interesting, like you said, about how valuable call transcripts can be and how like, okay, this can this can kind of feed in right. to like so many parts of the business. I just thought that was an interesting one where it's like, this is kind of a trickle right now. And maybe again, there's probably a better way I can try and, you know, once a month categorize this, put customer, you know, ARR

Jay Nathan: Yeah.

Jeff Breunsbach: against it and and whatever, make it really consumable for them. But at least right now, I think it's just interesting, like, hey, jump on the channel once a day and just see what customers are saying and like, you know, it takes you a minute to To breeze through these categories because there's you know two quotes under each one. And at least you get a sense of like what are what are our customers actually saying on the front lines.

Jay Nathan: Totally. Love it, man. All right. I think that's it.

Jeff Breunsbach: Cool. Yeah. Good episode.

Jay Nathan: yeah, cool. talk to you soon. See

Jeff Breunsbach: All right. We'll see soon.

Jay Nathan: ya.