Ben: Well, David, thank you so much for coming onto the podcast. David Steckel: thanks for one saying hello, meeting me and inviting me on. I appreciate it. Ben: Of course. let's start with a little bit about your background. Like what what made you enter polygrade? how did all of that kick off? David Steckel: It's the culmination of my career actually, which is fun. I've been about 20 years in the home service space. I was a custom home builder for many years. And then I started Setter, which was the first home management company. We raised capital from Nico Sequoia, from NFX, from some others, and expanded from Canada to the US. And then we were acquired by Thumbtack, which is a large marketplace in the US. And at Thumbtack, at the time, they were mostly focused on everything and with no actual focus. And so part of the acquisition was around how they could focus on the home and the concept of LTV. And so I helped pivot them there. But my core focus at Thumbtack was actually building out their partnership team. Like how can you grow through partnerships? And I also led the innovation team focusing on like, you know, improving the homeowner and the service provider experience. Thumbtack is now a marketplace for home services. It's a lead generation tool for service providers to get leads, and it's a single point of contact for homeowners to be introduced to service providers for work for around the home. So if you need your siding done, painting, anything in and around the home. Yeah. After that I was brought to Sears Home Services as chief product officer to rebuild their direct-to-consumer brand. Sears at one point was obviously a massive company. And after their bankruptcy, they sort of disappeared and all of the, you know, the the structure around the bankruptcy was coming to an end and they wanted to go back into their direct to consumer world. But when I got there, they were doing 10,000 jobs a day with their own employees, which was just shocking to me. that volume is insane. No one, no other company in the US has any volume even close to that. And Ben: Wow. David Steckel: the most surprising part was the vast majority of that volume was coming from warranty work. And so it could be coming from the OEMs themselves, like Samsung LG, could be coming from the extended warranty companies. Or the US has a something called home warranty, which is a bundle of service contracts for appliances, HVAC, et cetera, around the home. And as I looked more and more at the warranty space, I realized this was like a generation or two behind Insure Tech. and I had spent a bunch of time with insurance, et cetera. And then as I really pulled the thread, there is no service Titan. There is no guide wire. They never had a software moment. And the Ben: Wow. David Steckel: really funny thing was one of my first jobs at Sears was getting Sears off the mainframe. That's how antiquated these companies are. I actually thought it was still, it was just Sears, but when we reintegrated back with all of our customers, I was running into EDI files, not API files. these were very, very far behind companies. And that was sort of the genesis to Polygrade. And our initial name was actually ProPay, because we originally started to fix just the invoice processing and payments. But then we realized the payments issue was really just a symptom of the problem rather than the actual problem. And the problem being that the entire claims process, similar to what's been happening in insurance, and I've listened to a few of your podcasts right now, they're fantastic. the entire claims process needs to be rebuilt. And that's where Polygrade came from. Ben: Very interesting. you mentioned that there was no like guide wire, no big, big shop that was in this space. Like how do you explain that? David Steckel: So there's a couple shops in the space, to be clear, but they're not focused. So the interesting thing about Service Titan, for example, is they spent years only doing HVAC. Same thing, you have GuideWire, all they did was insurance. The tools Ben: Mm-hmm. David Steckel: that exist in the warranty space. They are typically FSM tools similar to a guide wire, not to a guide wire, to a service type or a house called pro, but specifically for third-party managed tasks. They don't do anything around the claim. It's really just work order management for pros that do warranty work. and the reason why, if you'll remember, which I'm sure you will, insurance had a behavior 20 years ago of building their own stuff only. And then Insure Tech came about about 15 years ago. And it took a long time, but now build versus buy, most often it makes more sense to buy now. Warranty never hit that moment in time. So everyone's been stitching together custom solutions on top of modules from different systems. And the issue with that is unlike a guidewire or a service titan that has A-class serious technology teams, these companies have IT departments. IT is typically about security, compliance, telephony, making sure the computers are set up. These these companies are, you know, I work with amazing technologists at the companies, but that's not their core competency. So but they had to do it. They had no choice. You needed to have some sort of software that could function to help you produce your outcomes. Ben: Interesting. And then what would you say is the switch from, you know, what was the status quo for a lot of the people using the technology you just just described to them now going to polygrade? David Steckel: So I would actually say there's a few. One, there's the macro, where everyone knows that AI is around. You have your board members talking about AI, you have your C-suite talking about AI. Like as chief product officer, I implemented a rule that X percent of code needed to be written by Copilot. So you have that from like a macro external. But Polygrade, we didn't start. looking for a we didn't start with AI looking for a problem, if that makes sense. We spent, Ben: Mm-hmm. David Steckel: you know, we took our 20 years of experience in the industry and knew exactly where the problems were. And a lot of the problems were in the software itself. And you're sort of talking about a paradigm shift that hasn't really happened yet because these Ben: Yeah. David Steckel: companies are still using bodies and spreadsheets to solve problems. And I I keep running into, we're gonna move to Salesforce. Who moves to Salesforce? Like in today, Ben: Mm-hmm. David Steckel: who who who makes that decision now? Ben: I I keep hearing from brokers that they're building their own CRMs now, so David Steckel: CRMs is one thing, but when you think about policy management and you think about the work order management and the claims management, it doesn't. A CRM is a different type of tool. But yeah, so that moment actually hasn't happened. We're the tip of the spear, and a lot of the a lot of the my current customers are just so hungry for change. And They're leaning in and you need to have executives with courage to make a decision to digital transformations don't typically work. And it gives PTSD to everybody in the company. So you have to know going in that there's gonna friction will be created. So I'd say my I'm very happy and I love my customers and they're courageous, but I'd say most of the in industry is fearful of change. Ben: And i when you say it's fearful of change, is this specific to To the claim space, or is this in in in general warranty? It because w when I listened to you explaining Polygrade, it I I was mixing two things together, I believe, right? So there's the whole home warranty space where the software technology to even run your system of record may not be, you know, may may be antiquated, right? May not be to up to the standard that something in SureTech now is, given that it started earlier. if a warranty needs to be exercised, then you have the software and you sort of built proprietary technology that is able to adequately assess whether the exercise is justified. Is is that a good way to think about it? David Steckel: Yeah, so we break down the claims process into first notice of loss. And then actually let me take a step back. There is a difference between warranty and insurance. Insurance doesn't want warranty be d to be called insurance because you don't want the regulators to be looking at them the same. Warranties are service contracts, really. And I'd say the core difference at the highest level is insurance is typically around a catastrophic loss. And so frequency is very low. And you have a loss ratio that's and keep me honest, a loss ratio is based on the life of the policy and the customer, typically. Warranty is different in that frequency is much higher. And so you're with a home warranty, you're actually expecting multiple transactions per policy in a year. So the focus of warranty is actually severity is their version of loss ratio, and they only consider at the job level. They don't actually think through the whole year at the policy level. So there's some fundamental changes Ben: interesting. David Steckel: there. And the other thing I'll say is warranty the majority of the companies in the home service space that I work with. They actually handle the repair, whereas insurance subs it out or writes a check. So because of that, you have a slightly different claims process. So you have first notice of loss, everyone has that. And then warranty is dispatching the technician. Some warranty companies will just send a check to the customer, but Ben: Mm-hmm. David Steckel: even still they have to validate does the customer, does the estimate of the pro, is that covered? So we actually get in the middle and between ethnol and dispatch, we do a triaging to understand what's actually wrong. Ben: Mm-hmm. David Steckel: Should you send someone, should you not? Because our focus there is how could we reduce severity? But we also want to improve the customer experience and reduce the total cycle time. So if you're not going to get someone there for two days, what can we do in those two days? And I've spent 20 years solving that. And then you have Ben: Of course. David Steckel: authorization, which is similar to insurance, like can you get more money? And then you have status updates, like how many times have you called an insurance company and like what's going on with the job? It's very hard to know. The adjudicator doesn't know what's happening with the job. And then so you have status updates, then you have invoice processing and payments. So we have a solution for each of those areas, but our core difference is we don't require you to replace all your systems. I know you're using a, like I said, a spaghetti of different tools and fragmented systems, but I don't care if you rip and replace yours. I'll just come in with a single module that's solving one of your pain points, and we can expand from there. Ben: Really interesting. I I have so many different directions I would love to take this in. I I think there's three things. So let's first talk a little bit more about warranty holistically, because I think you said a few things that were new to me. And then I'd love to know a little bit more about the processes you have in place. And then the final point, and also one of the reasons I wanted to chat to you was about how AI changes not just, the claims processing and and the things that you've already embedded, but also in general, you know, the willingness of your customers to build versus buy. So let's let's start on on the warranty side. Just for me to understand, if a company comes and says, hey, we have we have warranties that we sell, they're kind of service contracts, but we are fully anticipating that if a warranty has been deemed to be what do you call it, like exercisable, right? A warranty has been deemed to be active, that's a claim, yeah, an David Steckel: You mean a claim? Yeah. Ben: ethanol happens, right, and and we think that that's legitimate, then we'll just pay out money. Doesn't that then almost resemble an insurance contract. i if if they don't really have an intention to to to bundle it with a service that exercises the repair in some way. David Steckel: Yeah. So I just so oddly enough, I just broke my iPhone. I went right into Ben: no way. David Steckel: the corner of a desk and it broke my camera. And Ben: Sure. David Steckel: it took and my credit card gives me a warranty on my Ben: Yeah. David Steckel: iPhone. It took it's in it's insane, but it took AIG 14 days to then send me a confirmation because I had to send them I had to do all the information myself. I had to go get a quote, I had to send them all this stuff. And I followed up multiple times, got nothing. And then yesterday I got an email saying we're sending you a check. No one looked at my phone. Ben: Congrats. David Steckel: I none of so as it relates to electronics, I think you're right, but a service contract is really you're pay you have you've paid for a contract that promises service. So what they're validating is that you've found a legitimate provider to deliver the service. So I think there's a whereas insurance, and again, you're more of an expert than I am, they've already come up with a number for your property based on an actuarial model. And if you have a house that has been fully destroyed, they're gonna just write you a check for that amount. Ben: Yeah, I I I think thinking about who decides what how big the loss theoretically is is probably a good way to split this up. Whereby, you know, you you you've sent something to ARG and they sign it off, and a lot of the cut times in large insurance contracts, right, it's the insurance company who will go and and sort of seek out the easiest way to pay David Steckel: Is it though? It's it's my understanding that TPAs are very big and the insurance companies actually don't do any of that work. Yeah. Correct. Ben: Yes. No, exactly. The insurance company doesn't do it themselves, but they're on the hook to hire the TPA to go out and and and assess the property or whatever it might be. David Steckel: So that's the same. So we saw the original generation of a warranty was behaving the same with external TPAs, but these companies are even further behind than the. And so what you're seeing is a industry-wide move. And like we're I'm seeing it in auto, I'm seeing it in electronics. I'm seeing it everywhere. Everyone is now and this we could jump into AI too. Because severity is so high, is so high, and because frequency is so high, these companies are looking to themselves and they don't have the the luxury of having a reserve fund making money for five years before a claim. So they Ben: Cool. David Steckel: need to take hold of their operations immediately. And so they're all bringing the administration in-house. Maybe that is the why now movement that you were sort of asking about earlier. TPAs and warranty are typically focused on the regulatory and compliance aspects of the underwriting, but they say they do everything. And my personal belief is no one's good at everything. So the TPAs that are great at the legal and compliance side, they are terrible at the customer support and service provider side. And so you're seeing all of these companies bring the administration in-house. to save the money, prov own the experience, so to speak. Ben: It's funny you mentioned your iPhone, right? Because presumably there's a huge warranty problem happening now with the new duo release. I don't know whether you saw that with David Steckel: yeah. So that's interesting. So that's actually an interesting distinction. So you have there's three types of warranty at the highest level. And what's actually funny at Setter, my pitch was: you know, even the shirt you're wearing has operating instructions on the back on the tag, but Ben: Yeah. David Steckel: you buy a home which is the most complex grouping of assets you're ever gonna touch, and there's nothing, there's no manual to home ownership. But with warranty, it's it's kind of the same thing. Your retailer or your OEM must provide some sort of manufacturer warranty for a period of time. I don't know how Apple gets away with their warranty being as small as it is. But the duo folks, if the screens are cracking, Apple's walked, I'm sure their OE I haven't looked at it, but I would bet their OEM warranty isn't covering it after a certain amount of time. Right? So That means Apple Care, which is not owned by Apple, it's a TPA runs it. And I think Apple made like, was it $40 billion or something on it last year? Ben: Okay. David Steckel: Their company, and it might be AIG, is having to deal with this right now. And how many people bought Apple Care for the duo? And then did they underwrite properly from a severity and frequency standpoint? But the OEM Ben: Or did they just trust Apple does good good yeah, trust Apple builds good equipment? And as a result, the underwriting is done up front, which David Steckel: But imagine Apple's gonna get ahead of it though and say, even if you don't have the extended warranties, we're gonna fix this crack screen. It's embarrassing. Yeah. Ben: I would s I would yeah, I I would definitely assume so. Okay, that's that's interesting. So th then I think I understand that a little bit better. You you spoke about home services. So what is included in that? Is it like if if someone comes into my home and they prov David Steckel: Yeah, so Ben: they provide like a plumbing service, for example? David Steckel: Yeah, so again, so there's three types of warranty. There's the OEM manufacturer. So back well, let's move it back to home. So let's call LG. And then if something's every item you buy, it's actually OEM says these parts are covered for X period of time. And it's with appliances, it's like typically a year. HVAC, it gets longer. And then labor might be covered for a year, but very seldomly more than that. Then when you buy the unit, you're typically sold a service contract. And so that's the retailer's chance to make a ton of money. The OEM has typically nothing to do with the service contract. And so the service contract is like, hey, for the it's a brand new appliance, and then we're going to cover anything that happens, labor and materials, for this period of time if it matches the contract. Then you have home warranty, which is not about new built homes. It's about A bundle of service contracts that are underwritten blind. So the appliances in my house are all approaching, I'm nine, 10 years, so I'm approaching end of life for many of them. I could go get a home warranty right now. I'm in Canada, we don't have home warranty, but I could go get a home warranty right now, and I should. Because the home warranty companies they don't know anything about the what they're underwriting. Their actual model, actuarial model is a little bit more random than an insurance. Insurance. And so, but all three of them are responsible for the repair. So all three types of warranty companies either provide you with the contact number of a service provider, or they will dispatch the service provider. The reason they all dispatch themselves is because back to severity again, if they control the pro, they can control the pricing, Ben: Mm-hmm. David Steckel: which is the only way that they're gonna get to any level of margin. And their job is to facil facilitate the job being done. But the unfortunate thing is warranty cycle time to get something repaired with warranty is two to three times longer than if you called the pro directly. So did I answer your Ben: Interesting. David Steckel: question right there? Like the warranty company is responsible for the the repair and they typically Ben: I David Steckel: own the management of it. Ben: Yeah, you definitely answered my question. It just it it it throws a few more ones, right? Because the w when I think about I have appliances that might go and expire, or the the manufacturer's warranty might expire even if I bought an extended one. what do I do about that then? My intuition would be contents insurance, right? Or home insurance of some sort that covers that covers these appliances in case something breaks. David Steckel: Insurance doesn't cover breaking. Insurance covers loss. So insurance so there's something called equipment breakdown in insurance, which most people don't know about, but it it started On commercial equipment and it's boilers, but it does exist for homes. The deductible is very high, so I think use is very infrequent. And I've spoken to a few insurance companies, they do essentially no triaging and they're just writing the checks. That's their behavior. But insurance typically covers the loss around the failure of an appliance rather than the appliance. Does that make sense? So if the dishwasher's leaking, insurance doesn't r give you a new dishwasher, it fixes the floor. Warranty Ben: Yes. Yes. That that now now David Steckel: fixes the dishwasher or repairs the dishwasher, but doesn't fix the floor. So you Ben: Yes. That David Steckel: need to have insurance and warranty together. For true peace of mind. Ben: Okay. This was awesome and I love I I think I think we gave our listeners a little bit of a you know expedition into the difference between different types of insurance and warranty and super helpful. So so I appreciate that. For your for pot polygrade, what is the perfect customer? Like what what are the people that you're serving with this technology? David Steckel: So I s the warranty companies are my customers, the service providers to a degree, but we're the we work for the warranty companies. You could think of us as a very modern TPA. In we wanted to keep warrant the claims administration. So we companies that are looking to refactor the way that they handle claims, that want to provide a better customer experience, that want to make the jobs more profitable for pros, and want to increase their retention. And Also for their teams, their cost to serve, the tools that they're using today and the work that they've make their employees do, it's just mundane, repetitive tasks. Like make their jobs better. People that want to pivot their teams from support functions, which are cost centers, to success functions, which are revenue-generating functions. Those are our ideal customers. Not the ones that say everything's perfect the way it is. We're going home at 5 p.m. Ben: Really interesting. I wonder why someone would even say that. Presumably there's also like the the typical, you know, innovators' dilemma politics involved, right? Presumably if your platform is deployed really successfully, you you may actually need fewer people unless you're able to grow substantially, right? David Steckel: you do. It's it's clear. We've seen it every implementation. It is A reallocation of headcount is the direct outcome of what we're doing. You can't even grow into it. They can't, truthfully. So we do one of our one of our functions is we have, again, because we focus on severity, one of the leading parts of severity is the cost of the parts or the replacements. So we've seen all of our customers bringing the parts procurement and replacements in-house. So because they'll have bigger purchasing power than like a service provider. Ben: Course. David Steckel: We've we've automated that process almost entirely. Ben: Wow. David Steckel: Ca and so you would have I mean I'm try I I can't share any numbers, but you would have Dozens of people researching parts, going to different websites. We just forget that whole concept. If we know that there's a part, we double check does it fit the appliance. We double check during the triage what is the part that this is gonna need to repair. And if the service provider agrees, we'll just ship the part. So there is no that entire workflow is gone except we notify them. if something's fallen off the happy path. So if a delivery has failed, if the part supplier hasn't shipped the part, we actually notify the part supplier before we notify the customer. So we're actually helping them. So that's an example. You need to redeploy your resources on more profitable revenue generating activities rather than cost centers. That's that's actually a tricky part of what we're doing, and there's a lot of politics involved in that statement because managers are managing people. So if you move people from this vertical to another, you might have lost some of your team. nobody wants there to be riffs like reductions in force. So where are you gonna put the folks? but I don't like selling to optimization and headcount. It's just the inevitable outcome. Ben: Yeah. Yeah. David Steckel: And that's why we try to say you should be focusing the team on success activities. Like go Ben: Yes. David Steckel: retention and warranty sucks. Actually, here's a stat. Warranty Ben: Yeah. David Steckel: has 5% market penetration in its eligible market in the US. Ben: Well. David Steckel: 50% of that is given away for free by realtors. So at the purchase and sale, a realtor will give their customers a gift of a home warranty. So, what that means is Ben: That's fascinating. David Steckel: 97.5% of eligible households in the US choose not to have a warranty. And that's because warranty, we've all been pressured to purchase a warranty, and we've all felt that that's a scuzzy experience, and there is a bad reputation around warranty. So I'm trying to push my customers to investing more in the customer and pro experience. Rather than reduce headcount. Ben: Yes. A and and it makes sense, right? If you're reducing headcount, there's a limit to how much you can generate from inter a interacting with you and i if you're focused on success, right, the like sky's the limit, as a result, so David Steckel: Yeah, and we're giving the money back, right? We're shrinking the severity, we're shrinking the total time it takes to get jobs done, which improve the experience. We're shrinking the cost to serve per claim. There is money to invest back in the experience. Ben: Yeah. David Steckel: If I can't tell them what to invest in, I'm trying to encourage them to invest. And we're starting to see it. Ben: Interesting. I think when AI was starting to be deployed in enterprise two, three years ago, in in a in a greater sense, or the all the L new LLM technology was looked at, I think, early back then, there was an expectation that it would change an operations manager's job to to be more of an auditor rather than someone who actually executes. I think that was one way also to soften the blow, right? To say, hey, we're not gonna fully automate it. We still need you a human in the loop. How how do you see that having transformed your business? And and do you do you actually think that the long run might not be that far away where you could fully automate it and you get similar results? David Steckel: So I don't think there's ever gonna be an instance where it is there's always gonna be escalations. It's just what percent of total volume requires the human in the loop. Another way to look at it is today a company will understand how many inbound calls they get per claim, but what they don't know is how many outbound calls they have per claim. So if that like so you you'll have your team calling the pro, calling the customer, calling the part supplier. Like no one tracks that. So Ben: Mm-hmm. David Steckel: From a metric standpoint, you should be able to see inbound calls going down and outbound calls going down. Calls are still gonna need to happen because AI is not anywhere near the point where it could. And if you think about the calls that are required, if it's escalations, it's too complex for AI. AI cannot handle, especially voice. Voice is Ben: Mm-hmm. David Steckel: more deterministic than text. Ben: Mm-hmm. David Steckel: we've seen I mentioned we start with aggregating the data, then we start with, then we move to a copilot where it's like we're take we're saying you should this is what you should do, or we're flagging this, and then we automate. But there is always a human in the loop requirement for when it happens. It's just the sheer number of escalations go down. but an interesting thing is as we're like going through authorization, for example. We can say, hey, this customer has been with you for five years. And if your data was good enough, which it often isn't, we can say that they've never made a claim, but they've been Ben: Mm-hmm. David Steckel: paying you for five years. At that point, I would actually push that to your best salesperson to call up and say, hey, Ben, you've been with us for five years. Not only this claim isn't covered, or even this claim is covered, forget it. We're just sending you a new dishwasher. Thank you so much for being a customer. That's what I mean about success. I think you're gonna get more, you'll move more bodies into that role, which didn't exist before. but AI, what AI really changed operationally, and you know, my background is ops. A lot of software historically was structured data fields, right? Like it was all about how could you force fields for structured data input and output, and you had very rigid decision trees, rigid decision trees, and you had predefined workflows. You That never really worked. So any company that's implemented a Salesforce, for example, by the time it gets to the ops team, they they have their own spreadsheets and their own processes that are outside of the system. So you always needed humans to be doing the operations function because of that. And the second thing is, I saw this at Thumbtack, I saw this at Sears, and I see this at every one of my companies. It's nuts. If you the larger a company is, the less. And these companies don't even have product. Thumbtack obviously did, but they don't have the concept of product management. They might have something called product owner, which is project management. The business is never involved in the roadmap of the IT org. Which sounds ridiculous, but so IT is building things for them and for their structured data input and output, but not for operations. Ben: Well David Steckel: So coming back full circle to your point, this the difference now is that AI is non-deterministic. It starts producing the outcomes, like it's entering the data for you. So you no longer the type of software has changed and the UI of software is changing as well. So you no longer need people doing those things, but you have to move them somewhere else. Ben: into David Steckel: AI fundamentally changes that behavior through its ability to understand natural language versus decision tree. Context is what's getting most exciting across interactions. Whereas, how many times have you yelled at an IVR saying, I want an agent? And I've looked at thousands of transcripts that say that. Ben: Yeah. David Steckel: They want an agent because it's ambiguous and the agent's not helpful. It's taking too long. AI is at the point now where it can help that and it can make the decisions and execute most basic workflows. Ben: I wanna click one point deeper in something you just said earlier, which was that historically the needs of the operations teams weren't considered by the IT team for them to work on certain things. And I find that interesting because presumably that gave rise to things like SAS, because the internal operations team needed software off the shelf because their IT team wasn't serving their needs. I'm not sure whether that's how we understood it. Okay. David Steckel: So, no, so, but let's expand what you said with what I said earlier. IT teams are not technology teams. So, SETR was a technology Ben: I see. Okay. So David Steckel: company. When we got to Thumbtack, I learned about the concept of EPAD. So, engineering product analytics design. And then you have the pod structure. And I learned, SETIR, we had product management, Thumbtack had more. Product management lives at the intersection of the end user, the business, and technology. Even at Thumbtack, the pro in, I'll get in trouble for saying this, the product team was really not engaging with the business team. It was a product-led company. Ben: Yeah. David Steckel: And at Sears, it was the exact opposite. We had no product. It was just the business saying to the IT team, change this button. There's no why. What is the problem you're trying to solve? Every company I look I'm working with, it's the same thing. You have IT teams that don't have the background in product development, they have a background in JIRA and moving tickets along. They don't Ben: Yes. David Steckel: think about what is the problem they're trying to solve, they're more execution-oriented. So that's Ben: Sure. David Steckel: actually the vicious flywheel that this entire industry has been in. Is that you're to your point. They took SaaS off the shelves because enterprises are not technology companies. And Ben: Yeah. David Steckel: then you hire the implementation team from the technology company, you bring in a consultant to help you with your NetSuite Oracle. And then the consultant actually gets to stay with you on fractional for forever or Infosys. But the IT teams always had to exist from security compliance, setting up your phones and your computers. But then they everyone started asking IT teams to build things because none of these tools talk together. So that's where you have this tech debt. They're these are not technology companies. Ben: It's so fascinating because you would assume that if you look at the previous twenty years, some of those companies will have made the transition to, for the Sears online portal, hire an actual product manager who understands, user behavior, like what the you know, what the funnels look like, you know, what the clicks are, and then dec d d dependent on what the business needs and what the technology team can develop, builds the roadmap to to to you know increase the flow through the platform even more. David Steckel: Product management is a relatively new field as well, right? So you have decades of producing technology and software. Product management is relatively new, but people were just buying SaaS off the shelves. And so the marketing team, well, typically growth marketing is looking at the clicks, but you I've seen these roadmaps, man. Like there, there are no roadmaps with an e-pad structure or pod Ben: Yeah. David Steckel: structures. And what these companies don't understand is when you build something, you have to maintain it and then you have to improve it. So marketplace dynamics. You have supply team and a and a demand team. They must be separate. They can't be the same. Anyone who keeps them the same is going to screw up. And then you should have acquisition and retention. And ideally, those people are colleagues, but there's a bit of friction between acquisition and retention. Because if you're acquiring customers that don't retain, you're wasting our time. And if you're not Ben: Sure. David Steckel: retaining the right customers, why did we go spend so much money? Ben: Yeah. David Steckel: And then you have teams and pods that that's all they work on. Enterprises that in warranty, and I did work with some insurance, they don't have that concept at all. It's we're gonna change this button and move on to the next Jira ticket. And maybe that I shouldn't have said it that way, but it's not that's I think Ben: Really interesting. David Steckel: one of the core differences between functions of IT versus a CTO org. Ben: The core difference between the dinosaur companies and the new tech, upstarts that have only been found in the last ten, fifteen years. amazing. So let's move into sort of the the topic of AI and the future and how that landscape might shift. One reason I asked that question earlier is because if we know that AI can develop things much more easily, and it it you know, there may even be scope. if you're prompting it in the right way to turn someone who historically has been in IT into someone who looks more at it from a product lens, right? Because AI makes all of that a lot easier. what's to stop your existing customers to try and take what you've built and and do something that is maybe even more bespoke to their internal processes than what Polygrade can offer? David Steckel: So I think there's always gonna be some level of customization at the company level. I don't think that goes away. I think you want your software to be malleable. I think what we've seen or what let me take another example. We're I use Slack. And at Sears, pre-Slack, they built a version of Slack. My CTO built that. It was called Pebble. And but it doesn't make sense for someone to build Slack today. why would you build your own Slack today? Why would you build your own Microsoft suite? Why would you Okay. Ben: I I I don't know, I would I would I would challenge that a little bit. It may not make sense for you to build Slack with all the bells and whistles, like the the functions you can embed, the notifications that you get, like th that that is quite a complex undertaking. But to have a small team of, let's say, twenty people have an internal chat interface that you've coded and you save the twenty five dollars per month per person, at some point, right, that that th th the the the underlying mass do change. And so my my question would be, what is that point? does Slack now need to make make their licenses cheaper for David Steckel: Yeah. Ben: them to be more competitive? David Steckel: I think one, I think very few people actually look at fully loaded engineering costs. So I've yet to see when I talk about cost to serve, ever not not a single one of my customers has been able to tell me from a an IT standpoint how much they're investing in seats, maintenance, software, customization. No one's been able to tell me that. Ben: Mm-hmm. David Steckel: So I am a small company. I could build an internal chat tool tomorrow. And I could, you know what? I could save $200 a month or whatever it is. Why would that be worth it? Ben: Ha ha ha. David Steckel: So I think I think most people get excited about building something and it's not gonna be as good. It's actually gonna be more expensive from the initial cost and then the maintenance cost. And it doesn't improve over time. That being said, if there's something unique to your business, that's a whole different experience. But there's, I'm not gonna build a Google Meet. I like Google Meet. I'm not gonna build granola. There's no point. And it's this is an argument build versus buy. That is AI has made it a lot harder. Because you can produce MVPs with AI relatively quickly. at the enterprise level a lot of it breaks. because AI is not meant for the tech debt that's is handling the tech debt in all these systems. But I think you're moving a you're we're gonna see a move away from seat based pricing, where, you know, as the company scales, do I pay more? Outcome-based pricing, I've read a lot on it, with voice AI in particular. Outcome-based pricing is really struggling. And it's we're seeing at first it was call resolution, but then like was the call resolved? Like you you Ben: Yeah. David Steckel: can't have pricing have any sort of area of gray. It has to be black and white. So I think I think what we know is true is seat-based pricing will go will change. I think We do outcome-based pricing because we are so this is how many triages we did, this is the definition of a triage. But actually, as we scale and take over more of the business, we're I want to stop charging for software entirely, and I just take a payment processing fee. Once I hit that, I'm no longer charging for software. Anything. But Yeah, it was historically companies were selling tools and seats. AI helps you build the tools, monitor the tools, and removes the need for as many seats. So I think we're gonna see some sort of new combination what is the volume of your business, what are the SLAs? So we have an uptime requirement of ninety-nine point nine nine two percent. And if that goes beyond, I think it's three minutes, we actually have to start reimbursing money. So Ben: Yeah. David Steckel: there is a cost for me doing that and it's based on the volume, because if I go down for three minutes and there's three hundred thousand claims, that's gonna cost me a lot of money. So there is a a volume based assumption on that, if that makes sense. But I I'm seeing the whole outcome thing only will have to go away. It's gonna be there and there has to also be a base Ben: Yeah, d definitely a lot of changes. David Steckel: monthly fee I think has to come back for Ben: I David Steckel: maintenance. Ben: for sure. I w what what I'm struggling with a little bit is I I've heard the two years ago it was a lot around Well, AI is not very good at coding just yet, right? And then a year ago we see, okay, AI is good, but it's hard to it's hard to really have sequential deployment, right? You really need to have a developer that constantly uses AI, but you were just saying, you're asking your developers to use a certain percentage of AI AI to write code. And I'm I'm just like, I think now most developers it will be like 100. It's very rare that they even look into the code base. So that trajectory has really accelerated. David Steckel: Well they also can't even look at they can't understand the code anymore. AI code, I don't know if you've looked at it, and I'm not by any means an expert, but AI outputs code that looks different than human-generated code, so you actually need AI to review AI code. Ben: Yes. So so exactly. And so then my question is, if we take that to the logical conclusion and, you know, you're an entrepreneur, like the the whole your whole job is to think about what does the future look like and build towards that, then wouldn't we be able to have someone like Claude in their newest, newest model, also fully scan enterprise environments, infrastructure and make changes whenever that breaks. And and once AI gets to that level, I do think that the value of something like Slack that is Almost like a single use case tool or something like Salesforce, maybe less of a single use case tool, but like clearly still very nicely packaged, plummets, and something that is tying processes to one another, right? Something that basically allows you to really solve this process and then this and then this and then this, right? SAS plays. You don't think that's true? David Steckel: Well, we are that middleware. we're the connective Ben: Exactly. David Steckel: tissue between all these systems. Ben: Yeah. David Steckel: there's six systems in a company, and we're saying don't rip and replace those, because we'll connect them all. But the ultimate end is we become the surface layer, and then there's a good chance we'll remove those different systems down the road. So I agree with you. it's hard for me to take the contrarian arguments. But I think I think there's another part with AI is it's about the workflows and really deep subject matter expertise. I've said you can't why combinator yourself into warranty. And I stand by that because the operational complexity of the outcomes we deliver is so serious. Like Ben: Yes. we spent the first twenty five minutes of this podcast just talking about that, so f for sure, right? It's David Steckel: Yeah, so I think companies that are hyper vertical specific win. I think when you're talking about like one task companies, I think they need to evolve. you know, Notion is a good example of how Notion is doing the calendar, it's doing it's doing the transcription, it's doing its wiki, it's it keeps adding more. Ben: Email. Yeah, yeah. Mm-hmm. David Steckel: I think I mean I love Slack Ben: Trying to become David Steckel: compared to Teams, but And it's more of a cultural thing, I actually think, for me, than it is like a Ben: For sure. David Steckel: but I see companies will behave Notion or HubSpot's another good example. It's doing a lot. So you have to go horizontally for some of those companies. And then companies like me, I must be so sp vertical specific, but then to your point become that middleware internally horizontally. Ben: Amazing, that's such a good point to end on. David, thank you so much for your time. This has been this has been like a a huge learning moment for me here. David Steckel: Well, you made me think a lot, actually. Now I'm thinking about systems of record in general here. But Yeah, no, it's been an absolute pleasure. on. And yeah, it's a really great dialogue. Thank you. Ben: Thank you very much. Have a good one, David.