Latest / The Chief Customer Officer Podcast / EP010: You Can't AI Your Way Out of Bad Data

Transcript

EP010: You Can't AI Your Way Out of Bad Data · The Chief Customer Officer Podcast

• Jeff Breunsbach: All right, welcome back to another episode of Chief Customer Officer. It is Thursday, March 19th that we're coming to you live. Well, it's going to release on Thursday 19th. Yeah, yeah, yeah, yeah. All right, Jay, what's going on? How was your weekend? Jay Nathan: ⁓ yeah, you're right. You're right. I'm sorry. I was like, ⁓ weekend was good. My wife's birthday. Pulled that off. We went to a place called our kitchen. Have you heard of this? Jeff Breunsbach: Mm-hmm. It's awesome. We love that place. Jay Nathan: We didn't. mean, did you go to the one in West Ashley or in downtown Charleston? Jeff Breunsbach: Was it good? ⁓ I've been to the one downtown Charleston. Jay Nathan: Okay, a little underwhelmed. Little underwhelmed. Jeff Breunsbach: Okay, with the food or with the scene or with the whole concept? Jay Nathan: The whole concept was cool. The food was like 50 50. There was no like element to it at all. And then the price was just like, it blew me away. I we took six people. So it was like, we ⁓ Yeah, they did. Yeah, and then two of them were very mediocre. Like ⁓ ⁓ Jeff Breunsbach: weird wait so they did they not like bring out the dishes and explain the whole like all the all the ingredients all the stuff ⁓ Okay, man. Jay Nathan: can get into this more later, but one of the guys said, Hey, I've been saving scraps of meat for two weeks to make the shepherd's pie. And my wife was like, what? touched it at the table. ⁓ I ate all mine. It was the presentation was not anyway. So the concepts of people though is like, you don't, there's no menu. You walk into this place and they serve you what they've made that day. Five courses, small courts. ⁓ Jeff Breunsbach: That's kind of a weird thing to say. Alright. Yeah. They basically have two seatings a night. So there's a limited amount of tables and, know, we've, yeah, we've, we, ⁓ that's funny. We've, we, we've gone two or three times now and each time we felt, yeah, we, we've liked it with friends and like, you know, brought some wine and like sat down ate and stuff. ⁓ Yeah. Jay Nathan: and you get what you get. It's like a house, very small. Really? You can bring your own wine. Okay. Yeah. Cause even yeah, there's no servers though. So it's you get the chefs and the chefs are like, I think we got like a can of beer and a salt. Yeah. Jeff Breunsbach: Yes. Grizzly, yeah, yeah. can see that part of it for sure. ⁓ But well, ⁓ tough scene for your wife's birthday. You'll have to make that up somehow. Jay Nathan: All right. Yeah. Yeah. Yeah, I know. I know. But luckily she picked it. didn't. So said she wanted to get his, I didn't get in trouble for ⁓ Jeff Breunsbach: It's your own fault. That's cool. Well, trying to think what I did this weekend. It's two. ⁓ actually, I spent about a day and a half outside away from my computer screen doing yard work. And ⁓ let me you, it was ⁓ kind of kind of, you know, like manual labor. ⁓ I was like, you know, using a chainsaw, you know, ⁓ I don't know, ⁓ moving some dirt the backyard. And like, it just reminded me that like, Jay Nathan: ⁓ nice. you Jeff Breunsbach: you know, I should probably get outside more often on the weekends. You ⁓ know, I spend all my time in front of a computer screen. I don't need to do it on the weekends. It was super fun. We had my kids out there and like Caleb, know, he had the gloves on and was using like a little mini shovel and thought he was like so cool helping daddy. ⁓ we had a good time. But it was manual labor and like we hung out with some friends. Went to, down at Freshfields, they had like a bourbon barbecue bluegrass type of festival thing with like some tents. Jay Nathan: Yeah. Yeah. Yeah. Jeff Breunsbach: And so could set up your own chairs until you took the kids down there with my parents and listened like live bluegrass music and got some sun. It was like 75 out on the weekend. It was so nice. Yeah. Jay Nathan: Oh, it's beautiful. Yeah. Now it's in the thirties. Jackie from Clemson last weekend too. So he, we've got him home first spring break, which yeah. Jeff Breunsbach: Nice. All right, well, I think, you know, we're both in the throes of doing a bunch of stuff and getting things done. And so I think one thing that has been pretty interesting for me here at Junction is I say like in my past lives and even, you know, during our customer imperative days, one of the things that we've kind of always seen is ⁓ how customer customer teams and then businesses in general struggle to get a handle on, I'll call it just customer data, right? The ability for kind of a singular view of the customer that actually encapsulates like all the things that are happening, not just a singular department, right? So support has support tickets and they can see that and CS has a bunch of emails and they can see that and sales team has deal records and stuff, but it's all felt very disparate and separated. so... Jay Nathan: Yep. Jeff Breunsbach: was explaining to you, think one thing that's been kind of an advantage for me right now is we have a head of data and AI here. ⁓ And been a very big proponent of going around to each of the stakeholders in the business and basically understanding like what tools are you using? How are you interacting with customers? then basically funneling all that stuff back into ⁓ like BigQuery Metabase, which are like ⁓ the tools we use. And so now like I guess like where the advantage is coming for me and why this might be an interesting discussion is just that I'm going to set up some tools that we have plan hat in particular is like our CSP tool that we're using. And so now I've actually got places to go pull this data in and it's already structured from these sources. It's got connectors in terms of like UUIDs and making sure that we can tie it back to specific records. And so like, you know, I'm starting to think about, okay, how does that. help us enhance the picture? How does that help us also enhance some of the AI motions or AI agents that we can build? So a couple ones that are just coming to mind right now, this is kind of like the next phase of like what I'm building in, support tickets, you know, getting, I would say, not only seeing the support tickets, but I think building some agents around those to help the CSMs understand if like, I mean, I don't want the the CSMs dealing with support tickets, but at the same time, we should know if a ticket's been open for a long period of time, then we should have an agent that's telling us why it has been, what's the most recent interactions, and being able to pull that back to the CSMs so that they have, you're walking into your next call, you should be able to understand where these tickets are and why. And you used to have to go ask the support person to do it, so that's one. The second one is around ⁓ invoice data, and so understanding if they paid their invoices. And we actually have all of our invoice data coming from a tool called Sequence maybe, actually I think that's wrong. I forget what the tool is right now. But like another one is, know, can I help summarize for the CSM? Like where does this customer stand in terms of like their financials for us as a business? Have they paid their bills? How, you know, which ones are outdated? But again, could I actually go be using agents to ping and kind of bounce these things off to like surface it for the CSM at the right time? So those are two examples, but like, I guess ⁓ bigger underlying piece for me is the ⁓ ability to pull that stuff together because we've I guess gone through some of the pain of like making sure that this is all pushed back into a ⁓ singular or singular place, like, you know, tables and databases that we can, we can connect all together. Jay Nathan: Yeah, I think the as I think about like the maturity, because so I've had three conversations in the past few days with different companies, small to large, like one was 150 person company. And then on the other end of it, it was a 3500 person company that had been, you know, multiple acquisitions, you know, put together and just creates a very messy environment. Right. As you know, we've experienced that firsthand. the story is exactly the same. It's like we don't even know who our customer is. So when I think about this, like step one, if you're a multi-product or multi-platform or a company done a bunch of acquisitions or been part of acquisitions, the first thing to do is get a customer master defined. ⁓ What your master customer record? ⁓ In one the companies that I talked to, they've got six or different Salesforce instances that aren't merged, right? There's two that are merged now and it's a starting point. But if you look at the roadmap for that project, it's even months or quarters. It's like years, right? And because Salesforce is very tricky to get consolidated, that's the right answer, right? Ultimately to get everything into one Salesforce environment. But I think there's a ⁓ way to do Like you cannot afford to wait that long. If you're the chief customer officer, you're gonna be gone in two or three years. I can almost guarantee it, right? So if you're gonna make any progress at all, I think you have to have this data layer outside of your CRM or your CRMs and I'll put customer success platforms in that category because as you know, my perspective on this is that a CSP is a shadow CRM system. But you've got to have a customer master record outside of it. And then you start attaching elements to that. Right? Like you just described, you got to a common ID. So I know who the customer is, product data coming straight from the product relates to that customer with telemetry data coming from a tool like Pendo relates to that customer. What invoice data, what support data it's all got to map back. Like this is just fundamental. I can't imagine, you know, how we got here as a sort of an industry because ⁓ think it's more companies have this mess on their hands, then don't, by far. So anyway, I there's stages to it, right? ⁓ stage one is, can I get a customer master file where I can at least have a database of all my customers deduplicated, I know who they are, and then maybe even pointers to the other systems where ⁓ data resides, right? Because it's all siloed. Every has a silo. ERP, ⁓ Jeff Breunsbach: Yeah. Jay Nathan: CRM, marketing, customer success, support, it's all silos. And that does you no good for AI, nothing. You can't use it. MCP is not the answer, I don't think. Jeff Breunsbach: Well, that's yeah. Well, that's the, think that was like the, like the, I think the thing that's been interesting to me is, you know, this feels like it's been a problem for years now, right? This idea of trying to create this data telemetry and, and I feel like now if you haven't fixed it, you're actually going to be significantly hindered because it feels like what you need for the future is context and to be able to pull all these things together in easier ways. And Jay Nathan: Yes. Yep. Jeff Breunsbach: Like that's what's going to help feed these tools, these agents, the things that you want to build on top using AI are all going to need to either pull from data and context or be able to feed into like main systems of record. And so if your data is not cleaned up, then it feels like getting to the AI pieces is like, you're going to be able to scratch the surface, but it's going to be almost like point solutions in the individual departments again. And you're kind of kind of almost like, you're going to take the problem and just like exacerbate it because, Cause now you've got a sales team that's using some AI tools only on top of sales data. And you've got a CSM that's only using on CS data, right? Like you've just got all these teams just basically using instances. I actually saw something that was pretty interesting that is kind of related to this, but somebody ⁓ the that like they have their whole, all their teams are using cloud code now. And the thing that they're realizing is that like the engineering has all their own setup, all their own MD files. Jay Nathan: Right, yeah. Jeff Breunsbach: for cloud code and the sales team has their own MD file. Now everyone's got all these MD files and they're like, we could have vastly different MD files that describe what our company does or what we're trying to accomplish. And so it's like, shit, so actually the problem, now you can start to see it gets like multi-tenanted as you go out. Jay Nathan: Okay. Yes. Yeah. Okay. So here's something I started working on this week that I think you might find interesting. First of all, let's define MD file for everybody gets marked down. Markdown is a human readable. It's a, it's a very simple formatting structure for text based documents that you could open that document up and read it. But then if you paste it into a Google doc, for example, as Markdown, it would paste it in. It would have all the formatting, like the headings and the and the line breaks and everything would be in there. So what I started doing this week, I had the same thought that you did, right? Which is for my company, working on our go to market strategy right now, our offerings, who is our ICP, who like the personas within that ideal customer record. I leaning this idea of company as code. Have you heard of this? ⁓ Jeff Breunsbach: Nah. Jay Nathan: as a company gets documented as a markdown file there's one canonical version of it or ⁓ ⁓ folder structure. That's a that's a that is the canonical version of those things. And so we started creating our personas for outreach in we're to storm and GitHub. GitHub is a code repository, right? Because guess what? So our website is built on cloud code with cloud code. It's just HTML pages, but it's now in the same place ⁓ as Jeff Breunsbach: Yeah. Jay Nathan: those personas. When I started doing targeted outreach, I've built another markdown file that is, Hey, here's the style that I expect when we do outreach. Right. And I, and I had Claude build that for me after I had it do a bunch of emails. And then I coached it and trained it on what I wanted that those emails to sound like and feel like, and like the value I wanted to be in those emails for the recipients. So now I've got personas. I've got the outreach styles. All that, all that I need to add now is who are the people I need to go identify the people on LinkedIn. Here's the profile. Here's the persona that it matches up with. Here's the outreach style. Now all of a sudden, Claude can actually do something with that and go craft me an email that is actually perfect for that person as good as it can be. And it's hyper personalized. And here's the other thing. So I've mentioned Jack's home from college yesterday. He spent the day building. was like, wait a minute. we don't need to send a generic message to everybody anymore. And by the way, we don't even need send a generic email to everybody anymore. We can send a specific email to everybody based on who they are, what they do, what persona they fall into, what their company is up to. And not only that, why not just build a custom webpage for that person, a custom landing page for that individual? So we started, he built that for me yesterday. And so now we can actually generate with all three of those pieces, profile, the outreach template and outreach guidance in the profile we can actually generate a personalized landing page for that person. And then it's interactive too. So we get some information back from them. So it's, you're starting a two way conversation. Anyway, like you could do this with your customers too, right? That's the other, you don't have to be prospects. You can do it with customers. ⁓ Jeff Breunsbach: Yeah. Yeah, you could also like see how this also if I already know that the person is right, then I don't need to get information and you can freely give them value on those pages, right? Hey, download these things. Here's like, like you said, you can make them personalized. Here's these case studies or use cases or industry studies that you know, you are like super relevant to you. Two things interesting. One is I want to say it's Clay. I'm pretty sure that the marketing team at Clay did this maybe about six months ago. They kind of came out with like a Jay Nathan: 100%. Jeff Breunsbach: Hey, we've been building personalized landing pages. and so I think like they're, you know, I think theirs is probably built of like AI within clay and that's how they did it. And then like they fed variables from clay into like the final, I don't know, there's some way, but like, so like, I think you're right that like, this is the direction that people are going for sure. Like the leading companies are basically like, we should never be sending anything generic ever again is kind of the approach that I'm feeling. The second thing, and I think going back to your point too, so like the Jay Nathan: I'll be. Jeff Breunsbach: It sounds like the power of using GitHub as a repository is essentially now you've got singular source of these markdown files and you've got like, but then like anybody in the company can basically go make updates. They can kind of check them out, make updates and then it should feed all, you know, back into like all the other ones that people are using. again, like I think getting back to this idea that, that you, you know, all should be kind of singing from the same concert ⁓ in like the same music. Jay Nathan: They can make a pull request. Yeah. Yeah, that's the key right there is you think about it, ⁓ maybe the marketing team and the go to market teams, you plural did define personas, right? And we've been very clear about that. It's documented somewhere, but is it document like then then the customer success team has to pick it up and read that and interpret it and do their own thing with it. And the product team has to pick that up and read it and do their own thing with it. before you know it, like you're way off, right? Because everybody's got their own interpretation of maybe what was a shared definition of the person you're trying to, of your customer. I mean, the cool thing about what these platforms are now enabling us to do is like, we're just, like you said, working off of one, ⁓ HEM sheet, right? So yeah, it's ⁓ if you really let your mind start going on this stuff, And you, and you started the year like, well, why can't we personalize? Well, know, why can't we ⁓ it one by one? Well, why can't we just give them a webpage? Why couldn't I just build them a custom application? Like can do this stuff. All it just costs tokens, right? And tokens are ⁓ dropping in price. Jeff Breunsbach: Yeah. Yeah, there's a another quote I saw from from somebody that said basically, these days, having the meeting to like figure out if we should build something is actually more costly than just building the thing itself. Right? Like we might as well just go build it and then like, come back and present it be like, Hey, I built this is this right or you know, like, it's, it's almost like you should do it after the fact, which I thought was like an interesting perspective to Jay Nathan: Yes. Yeah. Well, you know, let me say one more thing about meetings because that that actually is what I was going to add a minute ago. They are the linchpin for getting this right. Because what you're doing is you're talking to somebody like all those. I forgot to mention this part, the personas that I built, the markdown files, all they came from. I had Claude generate those based on transcripts of calls with those people, those real people in the real world. And every time I have another one, I'm like, okay, here's another transcript. Go update the persona file now. And it's like, ⁓ okay, here's something we didn't see before, but we've added it to the market. So like those conversations are, that's why that's the human interaction piece, but it's so much easier than it used to be. And so much faster than it used to be, because we take that and process it immediately into our core definitions. Jeff Breunsbach: Yeah. So I think maybe like on the devil's advocate side of me, what I'm sitting here and thinking is like, you know, there's gotta be a downside, there's gotta be a risk. And I imagine the risk and the downside to like pursuing this type of strategy is that, like, how are you going to maintain this, right? Like, how are gonna maintain like the kind of the integrity, right? So that these, if I'm building custom landing pages that works appropriately, ⁓ know, like that's. Part of the allure of some of these platforms, right? The CRMs, the ERPs, is like they've gone through and had these kind of battle scars. They've figured out the kind of the necessary things to get right in order for these systems to be reliable and dependent and have all this kind of, got platforms and you know, they have ecosystems that they've built around them. And so as you start thinking about going down this path, like what is the, I guess, is the, do you see like the main risk as, okay, every so often I might have to take ⁓ time to. Jay Nathan: Yeah. Jeff Breunsbach: troubleshooting issue directly rather than having it within a system that is ⁓ inherently doing that already? Is that part of the risk? Jay Nathan: What mean by troubleshoot? Like what do mean by Jeff Breunsbach: Like if you let's just say you built a landing page for let's just say run that through run that example through right like what if you okay I said you know Jeff works adjunction but I sent him his competitor landing page because of something that went wrong in the code right and also now I've lost a conversation with the customer with a prospect because I sent him the wrong landing page and so then like how are going to figure out that that actually happened a and then B how are you going to fix it and troubleshoot what that issue is and then you know. Jay Nathan: Yeah. Jeff Breunsbach: Or maybe you built me a custom application and I click the button and the application doesn't actually work. And you know, now you've sent me something. I just, you know, like the devil's advocate side of me is like, okay, there's got to be some version of, you know, percentage of, Hey, this is not going to be right. A hundred percent of the time. But, you know, there's probably some percentage that you're willing to take and say, Hey, you know, if it's 10 % wrong, I'm kind of willing to take that percentage right now. Cause that means nine times out of 10, it's right. Jay Nathan: Yeah. I don't think you're going to send the wrong thing to the wrong person 10 % of the time. think more likely to happen is you're going to, you're going to come up with some copy on that customized personalized page. It does just not really relevant to the person. That's probably the bigger risk, right? It's like we missed a little bit there, but part of the part what we've got to do. And I'm talking about prospecting now, right? But again, I like I would encourage people to think about how this applies to the customer success side of what we do. Jeff Breunsbach: Okay. Yeah. Yeah. Jay Nathan: I can think of some ways maybe we talk about, like really actually have limited information. We don't have as much information on the people that we're trying to reach out to as we think, right? Just because they have a LinkedIn profile that says this title and like even you may ⁓ use words to describe their situation that they would use. And so you're not on the same page inherently, right? Even if you do sort of get like the gist of it, right? So I think, so I think this whole idea of thinking through the personalization of how you engage with people is to drive for two way engagement, right? You're not just trying to spray a message at them and get them to hold a meeting with you so you can demo, right? You're actually trying to make sure that you understand them fully and that they fit. I actually don't want to talk. Well, I shouldn't say this. I will, but I mean, in a, in sales setting, right? And to go to market setting, you actually want to avoid talking to people that aren't in your ideal client profile because it, it burns cycles, right? It burns time. You don't really have that much time. You wouldn't be talking to people who are a fit for your product as often as you possibly can. So like why not offer as much value as you can get into any message you send, right? You and I have always talked about that. That's a, you know, first principle for us. And then you know, use it as a way to drive engagement. So the pages that Jack built, they actually have three or four questions and it's like, you know, click this box, this box or this box. And then it edits like gives you the next layer of boxes. So it's like a little survey, but it's very simple, right? You just sort of click through it. It's like, okay, now I better understand what your problem is. If you took time and actually read those boxes and click the right ones. And then now let's schedule a call. If it makes sense for you, it's sort of how we get about it. Anyway, I went down a rabbit hole there, but Jeff Breunsbach: Yeah. Jay Nathan: I think that's the bigger risk than doing it, than actually sending something wrong to somebody. Jeff Breunsbach: ⁓ and I think to your point about like not knowing as much as you think about people, I think is, is like largely true. And I think like a lot of these data companies have like, tried to fool you over time of like, you can go pull all this data in, like you said, right? Like it's times out of 10, you get to there and you're like, okay, cool. I, you've just allowed me to figure out that company has 300 people, but that doesn't really, that's not a great, that's just a public figure. It's not really a marker, ⁓ one thing I have, but one thing I have been doing Jay Nathan: doesn't mean anything. Jeff Breunsbach: Yeah, which I'm curious if you so I've been using Claude Co work in a way where because a lot of companies that we deal with right now are private private companies. But what I've been doing is asking if Claude Co work can go research these companies and figure out if there are companies that are similar to them that are public or similar to them that have come out with public figures and documentation. And so then it'll go say hey, Company A looks most similar to Company B and here's the public information that I found about Company B and like. that at least gives me something to bring back to company A that says, hey, are you similar to them? You know, it seems like you are. And like, here's some things that we learned about them. Are these, are these things similar to things that you're thinking about for your business? And at least it gives you, like you said, like, I think nine times out of 10, you know, the customers that you're interacting with, or the prospects like are really just looking for, you know, edges on competitors, edges on the market, understanding about like what's happening and for a partner that's bringing perspective, right? Like, okay. Maybe, like you are my customer and we know I have a lot of information about you about how you're using our product. But like, I still need to know how you see the market and like where you think the market is going and where you're going as a business. so like that to me, if I can put those lenses together, it makes me a much more valuable partner to them. Jay Nathan: It's a great example. It like, if you want to know what Pindo's executive team is thinking, you should probably look at amplitude, right? It's really one of the only public cops that, in their world. Right. And so that that's a, that's a really good point. ⁓ Jeff Breunsbach: Yeah. And even like you could get, could take that Pendo example, right? Like you could even cause that last year, does that last year don't walk me or not? make that up. Service now. SIP, but you could even go say, Hey, can you go research? Hey, go look in these public filings about, know, this tool walk me and tell me what, know, where do think the direction's going on things? You know, cause they probably have some version of something in their public filings, uh, which I actually just, I just saw, I don't know if you saw this too, but, uh, the sec is looking to move. Jay Nathan: No, ServiceNow does. No, no, no, no, SAP does. Sorry. Yeah. You would hope. Jeff Breunsbach: quarterly filings not be required. And you could do it up to two times a year if you want, which, you know, I think in one respect, is it could be a good thing because then you get a little bit more longer term views on these companies versus just these short term cycles where they're just trying to, you know, this is not financial advice, but like, you know, sometimes you feel like they're trying to just pump revenue numbers up just to beat the quarterly earnings. And so maybe you get more of a longer term view, but ⁓ it'd definitely a shift because then you also don't have publicly available information that you could use to trade on these things and to understand. Jay Nathan: Yeah. Well, yeah, that's a whole different podcast probably. And I need to add myself better on it, but I've heard that. so let's talk about how you could use these same techniques to go engage your existing customers. Cause what I, ⁓ of the things you and I talk about a lot, Jeff, is that just because somebody's your existing customer doesn't mean they know who you are. They know anything about your product. They may have a negative taste in their mouth just because ⁓ They think your product is expensive for the value it provides or somebody said something negative about it, right? And really in the hallway or in a meeting. And so like those are the those are the people that are always elusive that we're trying to chase from a customer standpoint, right? You can use the same exact techniques to go get their attention. ⁓ again, it's not you're not trying to you're not trying to just ⁓ their attention so that they'll because they're not going to buy anything from you. If you're lucky, they're going to So you're not just trying to get them to take a meeting and do a demo, right? As you might be in a sales setting, you're actually really looking to engage them to understand what their perspective is on your product, maybe educate them on it. So idea that you could use ⁓ almost like an outbound campaign strategy to build your presence. So you're in understanding of what your company does from a customer success standpoint. That's really powerful too in the fact that that costs nothing now relative to what it might have cost before is a really important distinction. Jeff Breunsbach: Yeah, it almost, ⁓ so a couple of things come to mind. Like one, I actually put out a post this morning about ROI. Like I've been thinking a lot about ROI right now because we're kind of, in this cycle of, looking at our data and trying to understand like what should we, what should we be playing back for customers? Like what do we think it inherently means like value for our product and for our customers? And so I think to your point, like now, you know, is there something that's stopping us from going to actually We've talked about this for a long time, but personalizing ROI to those businesses, right? Like what's most important? What is the most important metric for you? And actually are we in the best place ever to go capture that, ⁓ from them and then be able to basically dynamically build dashboard or dynamically build something that actually relays that value back to them. I think that's probably like true. The second thing I think, is also really interesting and something that I've been thinking about the over the past week or so is, ⁓ upselling and cross-selling products. Like sometimes those are really hard. Jay Nathan: Right. Jeff Breunsbach: conversations, right? How do I go pitch this to a customer, like, you know, a current customer of ours? Are they happy? Have they had bad support interactions? Have they? And I think like sometimes we kind of over cycle it, but at the same time, like, now should we actually have a great view of the customer kind of where they stand, because I've got all this data and I can actually overlay AI on top of that basically summarize like, hey, where do think this customer is? Do you think this is the right time for them to be buying in our cycle? Like should they, you know, can it analyze things for us? Like, hey, they they're past onboarding, but they still haven't like achieved value on the first product, right? They're not ordering enough volume from us that would indicate that they're that we're in their ⁓ actual like flow of things. So it can help us analyze those things, but then could also help us understand like, yes, this is the right product to go pitch them. ⁓ here's how you'd pitch it to that, that particular customer So for instance, like we have a product now that we rolled out called external lab results, but essentially like we now allow you to bring PDFs of like lab results, and then we can essentially parse that, like using our own technology, because we've done so many of these blood tests, we now know what we can basically read PDFs. And like with, you know, accuracy, put this data into and digitize it for you so that your your customer who has a ⁓ PDF can send it to you. And now you can digitize that and put it into your dashboard or however you want to present that to a customer. And so to me, it's like a hugely valuable product, but like, You know, you kind of need to go understand that business. Do they have customers that are bringing PDF results to them today? Has this ever been brought up before? And so, like you said, can I go analyze? We have Fathom call recordings, we have emails, we have our Slack channels. Can I basically get Claude to go analyze all those things and figure out like who are the top 10 to 20 customers that we should take this to? And therefore like now I've taken, you know, my customer base down to 10 or 20 to even just go figure out what this motion looks like and the talk tracks and the early parts of this. Like, is that something that we can go do? Jay Nathan: Mm-hmm. Yeah. I mean, you just took it to like the nth degree me, ⁓ like to start simple, right? Like send a newsletter and your customers, Hey, like, did you know that our product provided this value? Forget ROI even, right? It's like, here's the value prop of what this product ⁓ for you over the past month. By the way, here are three other products and just make sure that message is short, sweet and good. Jeff Breunsbach: Yeah. Yeah. Singularity. Yep. Jay Nathan: to the right people that need to understand that, ⁓ we own that product. We own Junction. ⁓ know, they also have these other products. I didn't know that like, because remember, you're going to get like three seconds of somebody's brain. That's it. Yeah. ⁓ Jeff Breunsbach: Yeah. Yeah. Yeah. But it reminds me almost of like a, uh, uh, an idea of just like a better, or a different type of survey, right? But this should be based on almost like the contact record level and saying, okay, does Jay know about who we are? And like, are we getting signals back from Jay that basically he knows who we are? He understands the value, but like it changes maybe from like a single NPS score to like, you kind of break it out into different buckets of like, do they know who we are? Are they, do they think we're going to do? Do they have perceived value and do they know? Jay Nathan: That's right. Yes, yes. Jeff Breunsbach: Do they know all of our offerings? Almost like if you could answer those three things. Jay Nathan: It's a, kale based marketing. That's what it is. had one of the most, what are the coolest companies that, that I worked with when I was, an operating guy at JMI was, ⁓ this company, I won't say the name of it, but they had a marketing guy. They had, had this big platform, big, heavy, heavy platform for like big enterprise contact centers. Think like one to 3000 person contact centers where if you can optimize their processes just this much, you save tons of money, right? So they have very, very clear ROI, but they had this account-based marketing program. And by the way, not only did they use it for their new customers, they only had probably three or 400 targets globally. That was their entire TAM. So they had to be very, very careful and deliberate to say, okay, Jeff Breunsbach: Yeah. Jay Nathan: Who are the targets, these companies, who are the people, what roles are they in? And then they had this program, it was called Spreading the Flu, field level understanding of their product, FLU. It was such a brilliant, and they had these stages set up where it's like, hey, the first thing we're gonna do is we're gonna run this workshop and we're gonna infect everybody with the flu. It's probably a good, like everybody's got the flu right now, it's not a good analogy. But like it was such a good, Jeff Breunsbach: Yeah. Jay Nathan: program because it literally was dry. It was an account based marketing program and the goal was to educate at the individual level, not educating a company to your right. You got to make sure you know who the people are and you got to get them educated like for your top 20 % of customers. Let's face it. Maybe it's top 10 % that are are that are generating 80 % of your revenue right or 70 % of your revenue like we should be doing that. Jeff Breunsbach: Yeah, yeah. Jay Nathan: We should absolutely be doing that. And that's where things fall down with a lot of products. Either your product is not important enough to do that, ⁓ you don't know who all the people are that influence ⁓ buys your product. And so you gotta figure that out first. Jeff Breunsbach: Okay, well, I think you've given me my next task is to basically figure out if I could take cloud code and for our top 10 or 20 % of our customers, can we build some specific almost basically, could I take a QPR deck and build it into an interactive landing page that's like specific for them? And I think that would be a fun like outcome. Yeah. Yeah. And then Jay Nathan: Yes. Yep. And you campaign, how you campaign. Yep. And make it interactive too. Like make it so they can leave comments and interact with you right in there. I mean, as you start to peel the onion back on these things, you realize, Oh my God, okay, well that's a product anyway, right? That's something we could build and somebody might buy that and implement it for their organization. So like very quickly you're, you're, I don't know. I just spiral on these things. Jeff Breunsbach: Yeah, yeah. Yeah, yeah, for sure. Well, I know we've got a couple minutes left. Anything else top of mind that's like short and quick we can hit? Jay Nathan: I don't think so. Oh, we so we're doing an interesting webinar this this Thursday. By the time this comes out, it'll be noon today when people hear this noon Eastern time, but we are doing a webinar. Have I told you what the title of it is? I invited you didn't. Okay, the title is it's 2026. Do you need a customer success platform anymore? And we have we have a pretty big crowd registered ever like pushing 250 now. Jeff Breunsbach: Nah, nah, go for it. Yeah, yeah, you did, but yeah, tell me. Jay Nathan: for it. So it's interesting. It's actually attracted a bunch. I noticed a bunch of folks from the CSPs signed up to join. So clearly they're, they're curious about what's going to be said there, but it's not me talking on facilitating. We've got Ziv members, Ziv Pellid and Miranda to conski who are going to be doing all the talking there and share their experiences. So ⁓ really They come at it from, from both sides. Like Jeff Breunsbach: We go. Jay Nathan: Miranda's been a three or four time, you know, CSP customer. Ziv has built everything in house, right? And they, so they have two different ways of looking at it, but I think a lot of interesting content. had a prep session. was really good last week. So I'm excited about this one. Jeff Breunsbach: Cool. Yeah, I think the perspective will be interesting. So I feel like you can do it either way, but I feel like the, know, I think like as you think about where the future goes, right, it feels more and more that you've got control in your hands. And so how do ⁓ take that as a CS leader ⁓ as a business trying to build the perspective around the customer? So ⁓ that's cool. I like it. Jay Nathan: Yeah, I do have an opinion. I'll just put it this way. I don't need multiple CRMs. That just that that is the antithesis of everything we just talked about. And it's not doing us any good. So we've got to figure out how get away from that. Jeff Breunsbach: Go for it. Yeah. Jay Nathan: But that's not a new perspective on my part either. mean, we've you and I have talked about that for years. Jeff Breunsbach: Yeah. All right. Well, I think this is pretty fun just to talk through. It's cool to hear. I love that you got Jack working in the business, by the way. Feels like his skill set like sets up nicely for that. but yeah, I think we talked through, you know, a bunch of ideas just building around cloud code and the files I think are super interesting, trying to work from singular sources, ⁓ thinking data and how you're, you know, essentially going to start to be behind if you can't pull these things in one place. So, This is pretty fun. like spitballing this. I think I'm actually out next week. Shocking news. I have to go to a all day planning session for our company. So we'll have to figure out if we're going to do something for next week. But I think in that ⁓ it be fun to keep noodling. I was like taking specific things figuring out how would you adopt it for AI. So if I was doing QPRs ⁓ today, like does that mean using cloud code or AI? Like if I'm doing onboarding, what does that mean if I'm doing Jay Nathan: Couple years ago. Yeah. Jeff Breunsbach: But I think it'd be pretty interesting to think through some of these situations, just like your prospecting example, and basically figure out, okay, if I was to use AI, what does that spin look like? So I think it'd be fun to try and that in the future. Jay Nathan: Yeah, definitely. Alright, so you're out next week. I'm going to be at a conference, so we'll have to come up with something. For next Tuesday, we'll we'll figure it out. Alright. Thanks everybody. Jeff Breunsbach: Okay. Yeah, all right. All right, we'll see you soon.