Episode Transcript
Speaker 1 (00:00) A SaaS software development company called Cinch was able to scale its business by helping other companies manage their assets in the maintenance space. If you're in the marketing or even in the AI or customer service space, then listen up. Is your AI agent mishandling basic conversations or causing hallucinations? Speak to GPD Trainer. GPD Trainer fully manages your AI agent for you, eliminating hallucinations for good, while also drastically improving your AI agent's response. Go to gbttrainer.com to learn more. So our guests on the show today, I've got Dan Salas and also have Alfredo, who's the chief marketing officer. Dan, who's the CEO and founder of Cinch, an asset management SaaS platform company that helps high stakes industries for their maintenance and repair operations. They're based in the U S and this dynamic software company delivers a cloud-based platform that simplifies everything from preventative maintenance to inventory and vendor coordination. all in one flexible system powered by AI. Dan and his team, they partner with industries such as healthcare, aviation, manufacturing, and field services to keep equipment up and running while also being profitable. Hey Dan, hey Alfredo, welcome to show. How are y'all doing today? Speaker 2 (01:14) Hey, Kyle, thanks for having us. Speaker 1 (01:16) I'm glad to have you guys on board. Super excited to have this conversation, aren't you? Speaker 2 (01:21) Absolutely, yeah. I'll just tweak your introduction just a little bit because we're a US company, but we have a strong presence in Mexico. So actually, Alfredo is down in Guadalajara, I spend quite a bit of time. I'm in Colorado right now, but we spend quite a bit of time down in Guadalajara. Software engineer is down there. Did I lose everything? Did you lose me for a minute there? Looks like it froze for just a second. Anyway, we hire software engineers down there. And so that helps us bring the product to market for less money than you'll see our competitors for. And then the other thing is we're in this medical space and we're massively interested in all of Latin America. So that Guadalajara, a base for us is kind of where we launch off into Columbia and the rest of Latin America. And so we have a whole strategy for how we're going to be moving end of life equipment from the United States down through Mexico and into Latin America and so on and so forth. we're kind of, we're both, we're a United States corporation, but we do a lot in Latin America and love working with our Mexican team down Rock stars in Guadalajara Speaker 1 (02:32) No, thanks for it. That's great. mean, just to be able expand in different regions, I think that's a huge game changer for clients in different industries. So walk us through kind of the journey that led you to start Cinch and how did the company get birthed? Speaker 2 (02:47) Well, sure. So I'm back in quite a few years ago. I've had a server hanging off the internet since pretty much the inception of the internet. So think back in the Bolton Board days, before the actual launch of the internet. I was trading meat, actually, again, in Mexico in a place called Aguascalientes in Guadalajara in Mexico City. And we were trading meat between the United States and Mexico, and I was using bulletin boards to connect and talk to my customers. And then also using, I think it was Dbase 3 or something back then to manage that business, And then as I moved into other businesses after that, I was using access databases to manage all of my businesses, had various warehouses around the world and moving product around, so on and so forth. It grew out of that because the fact that there wasn't anything that I could run my entire business on. could run inventory and then do it and all that. I always used a homegrown solution. Then something happened in my life where I actually had to, I couldn't travel as much and I had to stay put in one place. I had to take care of my kids is what it was. When I did that, I said, well, I got to go completely online. I'm going to dive into this software. That's like the late 90s or something like that. And so we're doing consulting, right? And I already have all of this kind of infrastructure and stuff that I've built throughout the years, technology wise to manage a business on. And so I start incorporating that and rolling that out to these customers on a professional services kind of basis, start building this platform out, right? And so the answer in terms of kind of what took us from there to here and why it's actually taking so long is because we're playing very wide. So I've said that customers need a comprehensive solution to run everything on. That means their website, they can't use WordPress, and that means e-commerce, they can't use Shopify, they can't take all of these desperate systems and integrate them. And so I've been saying for the last 20 years or 25 years that they need one integrated solution. And that's what we've been pushing to get to. So we're trying to replace everything from SAP and Oracle, all the way through to Salesforce, WordPress. So today, our customers will host their entire website with us. So they don't need any other software. They don't need WordPress. They don't need CRM software. They'll host everything on our platform. And then I'll hear from investors and other people saying, well, aren't you playing really wide? That's really, really aggressive. But the thing that I've been hearing for the last 20 years from these customers, so let's look at our biomed customers, for example. So our biomeds are going into hospitals and repairing patient monitors and MRI machines and so on and so forth. So they need really comprehensive software around that MRO process, that maintenance and repair operations. so that they can report, the hospital can report information to the FDA of what the state of the equipment is in and make sure it's been inspected and all that. So there's a lot of software around that. Our competitors provide that software. Our customers say, well, we need that software, but we also need a website. And then we also need some e-commerce because we're selling some used equipment. And we also need some CRM and we need some coding software. And then we also need contracts. And so the list goes on and on and on. And I look at my, you know, investors out there that are saying you're playing too wide and kind of the people that think that we should be narrowing our focus and say, these guys are asking for it. They need it. They have to have everything and they can't integrate. So I've said for many years that the promise of interoperability is a failed promise. And so this whole idea back in the 2000s that everything's going to talk to each other and it's all going to work great has been a failed promise. And to prove that point, it's 2025. And if you walk into a Home Depot, I mean, you don't get it with Walmart because Walmart's huge, but Home Depot is pretty damn big. But you walk in here and say, you know, how many of these things do you have on the shelf? And they say, well, we're supposed to have 35, but I don't really know until we walk over there. So even Home Depot can't integrate all of these systems and make them work well. And so that's what's driven me to build this massive platform that'll handle everything. So we're doing credit card processing for our customers. We're doing this MRO piece. We're doing inventory, very complicated inventory with license plating, lot tracking and serialized inventory. All the way, it's like the full ERP suite. So that's the long answer how we get there. Speaker 1 (07:29) It's like initially when you first started off, was like you yourself were having all these different solutions that you had to go back and forth. And I imagine that that hurts like productivity because you're bouncing between applications and you get to the one application. You're like, what was I doing over here? So in this case, like some of your clients, instead of them having multiple avenues of processes and apps, they can almost just come to to cinch and it just keeps it all in the same space. So I think it's definitely helpful. what were, I guess, like when you started this, what were some of those biggest, bigger challenges you faced when bringing in some of this new SaaS, especially in the maintenance repair operations, like bringing those platforms in the market, what did you have to do and what did you have to navigate around some of those challenges? Speaker 2 (08:11) Well, it certainly is change management, right? Because the software is so dramatically different than what they're used to seeing today. So you'll have two types of customers in the field and two types of technicians, right? You'll have those that are just simply using paper, right? And then you'll have those that are using kind of an older and equated system and they're used to downloading everything in the Excel spreadsheets and then doing everything there. So we have this little inside joke because their customers are, they have these quote unquote data analysts. They're like, I know how to do spreadsheets. And they're like, can I get all this in a spreadsheet? And we're trying to go to the CEO of the company saying, you don't really need that. Just tell us what analytics you're trying to pull and we'll just write those for you for 500 bucks, whatever you need and throw it out. And then you can do 10 of those for 5K and you'll have everything you need or 20 of them for whatever. And you don't have to spend full time, you have somebody. And so for them to get used to something where you don't need that, where you don't need to download the data into a spreadsheet and run it analysis over there. They're used to downloading the data into a spreadsheet and then delivering that to like the Joint Commission. You don't need to do that now. The ones that get it, we'll see our little asset tags and it's really funny. They'll get these little QR codes here and they'll say, well, the Joint Commission wants to come in and spot check everything. We say, all you got to do is just walk up to a phone, tell the Joint Commission to scan that, and it'll tell them exactly what the state of the asset is in, and they're just blown away. Then you have the ones that are on paper. As a part of that change management, the other thing we run into is that we're dealing with that mid-market to smaller companies. In that mid-market space, they all think a platform exists where you can do everything. And so they'll be frustrated, well, why can't we do invoicing? expectation. Yeah, the expectation is this. And you're trying to explain to them that you'll get on these calls and will have all these requirements. OK, well, you would typically get SAP for the MRO. And then you're going to go and get Manhattan Associates for the inventory. And then you're going to go get something else for this, this, this. And it'll cost you about $2,000. four or five million dollars to get it all pieced together and take you a couple of years. And it still isn't gonna work all that well, right? And so it's that change management and getting people to understand that this is, but once they're on board, they're they're blown away. I mean, once they see how well everything works and how well it can work, they're... Speaker 1 (10:43) Cause you mentioned something Dan though, like that was interesting. Cause I mean, like even starting off on like on paper or whatever their process is. Right. And then to go from doing that for, you know, three, four, five, 10 years to going, Hey, look, we're going to put it all in this and it's going to be 10 times easier. I imagine there might've been like a little bit of like resistance initially, like, Oh, this is how I've always done it. But now I have to, how do you, how did you help them kind of, I guess, overcome that so that way they can, they can actually jump into your software. at cinch and get some success with it. Speaker 2 (11:15) Sure, so it takes, we need strong management with our partner, with our customers. They need to come down and that's one of the things that's taken a while for us to understand. So when we're in that discovery process with our customers, if we find out management isn't very strong, that's where we back away and say, this is not gonna be a good fit, this is probably going to fail. And it's the customers that say that, you I told my techs that they're going to use the software or they'll have to go get a job someplace else. That partnership and having our customers management team understand that when the tech says, well, I'm having a little bit of a problem here, I'm going to go write it. You think of it as a checker at Walmart using this and then saying, this thing didn't. this barcode didn't scan on this water. And instead of looking at, you know, I don't want to wait for the guy to go find it and all that. So I'm just going to write down a piece of paper and how much it was. And the guy at Walmart is like, no, no, no, no, you're not going to do that. Well, it's easier for me. I don't care what's easier for you. needs to go in the system because we have inventory and all this stuff behind it. And so that's again, that's been the kind of that major challenge of bringing this software to these. Speaker 1 (12:27) love the analogy too, like with the Walmart, well, this is easier for, at the time, yes, it may seem that way. Make that switch, that change management. Now you're on the right track to where it's not only gonna be easier for you, but it's gonna be easier for everyone down the road that's gonna have to be able to analyze it and get the business results from it. Speaker 2 (12:32) I guess we're Yeah, we just did a deal with surgical trays. Unbelievable how complicated it was. they, you know, they think of these surgical trays at a hospital and, you know, they're going through several of them for surgery, right? So you're to go in and do a knee replacement or something. The surgeon says, well, I need, you know, this one, this one, this one, and this one. They bring it all in and he rips them all apart and he goes through it all and uses some instruments, doesn't use other ones. Well, the whole thing, each surgical tray is 150 instruments on. It's got to go back through the whole process. And these instruments are worth three 100 to 1,500 sometimes upwards of $2,000 each. So a pair of tweezers for surgery, they're registered in the FDA database as a device, are going to be $300, right? And so they walk away and they have to be tracked. And the way they were doing it before on paper was just, it's unbelievable. We rolled out this, we rolled something out that is very, very quick for them to use. It's like three fields. Speaker 1 (13:36) imagine. Speaker 2 (13:43) Enter, Tab, Enter, Tab, Enter, boom, next field. And they're blowing through these 150, what used to take them hours, they're blowing through it in 15, 20 minutes. And then the reports that their customers are seeing and the Joint Commission and the regulators are seeing are like, wow, this is exactly what we need. Because they want to see things like, was there rust on it? If there was rust on it and you cleaned it off, How many, was it there the next time? And the third time, was there three times, get rid of that, get a new one in here. And so those commission, and so we're able to report on that much better. And with all that being digitized as opposed to being on page. Speaker 1 (14:21) I'm talking a little bit more about some of the clients that you're working with and you've seen some of the traction, especially when it comes to this day and age of AI powered tools. How have you been able to implement on the AI side tools that have been really just taking your business and even your client's success to the next level? Speaker 2 (14:42) Wow, so the AI tools have turned out to be massively better and moving along quicker than I had thought. And so we're using AI across the company in a lot of different ways. And more of it's in the backend than it is in the frontend for our customers at this point. Although we are using it customer facing in several ways, we have a huge advantage there because of our architecture. So our architecture takes a relational database and ties it in with a text-based search database. So think of Elastic or OpenSearch, right? And so because those are always in sync, we kind of already have our, you know, think of it as Snowflake. We already have our Snowflake instance there. You know, instantly the data is already there and we can run analysis against it. So you can go and, and update a procedure on a patient monitor, for example, and then go ask the AI agent, when was the last time this was updated? The AI agent knows it instantly because it has access to our data. That's one aspect. We're slowly rolling that out on the front end, but the interesting part is all in the backend. There's some change management there as well. You're going to our developers, was it three weeks ago that some, forgot who came out that said Jim and I was just. the best for coding, the new model. What is it right here? Yeah, the 2.5 Pro previews what we're using here. I went and I said, all right, let's go look at that. Awesome. Especially because we use Angular. it's a gemini Google product. Angular is a Google project. So we're in there working with Angular all day long. So I think that for your listeners and your business owners out there, and in terms of how you would use AI, I think you have to think of it in terms of, I have a three hour task. So let me frame it in terms of, ⁓ yeah, like let's say it's a development task. I wanna go create a form that can collect instruments on a surgical tray, right? Let's say that's gonna take 12, 15 hours, right? What I'm saying you need to be doing is you need to be spending probably 10 hours on AI and on the prompts, okay? Speaker 1 (16:29) breakdown. Speaker 2 (16:48) and developing those prompts to create that code and create all of those forms in the various database structure that you need. So you dial that in and then running that and cleaning that up. It could be the same thing in marketing. So let's say you have to go do a marketing task that's going to take you 15 hours. I'd say you really don't get into the task first, spend 10 hours on the AI, getting the AI, the agents and learning that. get your project out the door. And then the key is, and maybe you can help with this, Kyle, I don't know where you store the prompts. So we built a prompt repository on our site so we can know what those prompts are. So the next time I need to do that, next time we need to build a form, we go grab that prompt and say, here's our form builder prompt that we took 10 hours building. And we say, well, we don't need a form for surgical trades. We need a form for electrical outlets. Go change these things in the prompt. rerun it and now instead of the task taking us 12 or 15 hours it takes us two or three right and eventually we're trying to get that down to one right so we're using Speaker 1 (17:54) It's almost, it's almost like, I mean, just, I mean, it's with each different tasks. Like if it takes 12 hours for one, one task to be completed, well, you're, what you're saying here, Dan is like, Hey, look, spend the 10 hours on, on developing the AI, the prompt, the process, the systems. So that way when you run it again, right, you've got it all created, run that task through it, pumps it out automatically through AI versus having a team of three or four or five people having to go through it and spend hours. Speaker 2 (18:20) Yeah. There's some change management there with engineers. I get some pushback from the engineers. They're like, well, I wrote the prompt and it didn't do this, this, this, and this. It's like, because you forgot to tell it what database connection we have and you forgot to tell it about this. You need to tell it about all of these things and you forgot to tell it what version of Elasticsearch we're on, for example. The whole team's over the hump now. Once they got over that hump, they're like, oh, wow, this is great. And so don't think you're seeing, people are worried about losing jobs. This is not a job loser. This is making it more efficient. This is making software engineers' jobs more efficient. think, what I don't think you're going to see AI do, and I'm sure this is controversial. I don't think you're going see for hundreds of years. can't think. And so smart software engineers are like, well, how do I want to manage my surgical trades? What do I need to report on? What do I need to store the data? If I'm clearing rust off, does that need to go into my relational database so I can run a business rules engine against it and build a customer for that? Or can I just record it over here in another database so I can just report on it? And understanding what us as human beings want and what we need and what's important. are what a lot of software engineers are doing. They're thinking, I want to build this and I want to do this and I want to do that. The code gets in the way. The AI is getting the code out of the way. Another analogy I use there is think about movies. AI can kick out action movies all day long, Mission Impossible movies. that genre has been tried and tested in those. mean, can kick out these kind of action movies that have, Hollywood has been making movies where there's not a lot of dialogue, a lot of action, and they can export it to China, because that doesn't have to be a whole lot of dialogue. So you can create that with AI all day long, right? Tell me AI is going to create pulp fiction. Speaker 1 (20:27) you Speaker 2 (20:28) I don't think you need Quentin Tarantino to create Pulp Fiction. And so I think that that, know, and I think you're going to need a Quentin Tarantino for the next couple hundred years, maybe even beyond that. But is would it help him today to create the movie? Absolutely. You know, could the movie happen quicker? Could the thing could the, you know, the 20 or 50 movies he has in his head that he wants to get out? Could I help him, you know, produce those quicker? Yeah, and that's what we want. And all the things we have in our head, of things we want to make better in the world and things we want to do, can AI help with that? Certainly. But we're the ones that have to think of it. Speaker 1 (21:03) I love that too. Just like I've heard that phrase before, like it's, we're working with AI, not just have AI work for us. Like it's going to be a tool to assist almost like a car is to get us to a destination or a plane is we're just getting from one place to another. I love, like I've heard this term, like it's, it needs to satisfy like the last 20%. And that last 20 % is what makes going from a B to an A right to a home run just a little bit further. And I think that's where the human comes in where Hey, the human can do more than what the AI can. The AI might be able to produce a lot, but that fine tuning and the actual good detail that people truly care about, that's what's going to take them to that next level. And I want to shift gears just briefly. Alfredo, I know you're in here on the call here and talking about the AI. And I'm sure you've got your thoughts, how has it changed, especially on the marketing side, how has it changed your perspective and how you've been handling your workflow, both with your internal team and even with your customers? Speaker 3 (21:59) I've been cutting out some other tools. I remember I used to, after an expo, I'd have a stack of cards and I'd go through one by one updating cam card. But now with how good the optics are on the iPhone, you just take a standard picture of like 10 cards and make me a rough guideline with columns and create me a spreadsheet as to what I've captured. And then we play with that. We correlate it if we have it in our database. if we've before, we send a... a friendly follow-up email based on what we spoke to before. And if it's a net new person, we're clear about starting them off on a certain drip campaign around that. Definitely, I'm not gonna lie about it. I have used it for drafts and I have used it for editing and I have used it for translating. And yeah, it's been a good assistant, but not a replacer. At the end of the day, you know, our designers and our media specialists are coming up with some things, but at end of the day, Dan and I are looking over it to see if it's a message correct, if it's on target, of what we want to communicate and to whom. But yeah, it's a huge, huge game changer for sure. Speaker 2 (23:01) Kyle, there's something really interesting there to talk about as well. I'm launching something called IAMhuman. It's a project, think of it as an open source project, it's probably close to Creative Commons, right? OK. You understand Creative Commons. So similar to that, we need something similar to that for AI. Because when I send an email out, and I'm sure a lot of your listeners feel this way. Or when I get an email, I look at them like, dude, that's AI. Like, thanks. You're like, they go tell AI, well, what? write an email that says all, highlights all the requirements that I want to run a business on or something. Like get this email and it's got all the, this, this, this, this, this, like, dude, you haven't even thought about this. She just went to AI like, good luck getting all this. This is ridiculous. Like, I mean, and then when I write emails, you know, so I think you think about this and you probably feel the same way. When I write an email, I want to tell my listener, I want to tell that, that person, I didn't, I wrote this. AI didn't write this. I created this. I thought about it, it's not AI, and I think there's three levels. I wrote it completely, I wrote it with some help with AI, or this is just generally AI. And I think that moving forward, we need to have something that distinguishes that, and we need it before the regulators come in and screw it all up, right? The EU, United States, Congress, all these idiots are gonna screw it all up if we don't self-regulate. And so we need to have some sort of a notification on everything that, I'm a human. or I'm AI or I'm some sort of a mix. We get these calls now on AI and it's like, they need to tell you, think right up front, I'm AI, right? In fact, on your phone, you should be able to say, right, I don't wanna talk to AI. If you're not gonna call me, I don't wanna talk to AI, I'm not gonna, but I think that we need to, and I say that just cause Alfredo said, we use it for some writing, but I don't use it a lot. I read everything my customers send me, And I reply personally and write paragraphs about, know, if you're trying to manage this surgical trade, you want to do this and this, you should think about A, B, C, and D and how we do this and how we do that. We'll integrate this and, you know, all these very, and that's where I think AI can't get it. That's where AI is just this tool. Again, it's like saying, well, you know, Buggy drivers are go out of business because we have cars. No It's just gonna make it easier to get from point a to point B Like you said or airplanes are gonna make you know, we're not gonna need cars anymore We still need all that just makes us go there get there faster. But yeah, I think it's Speaker 1 (25:39) Right though like the different the differentiation like it needs to be clear and and it's so funny and I'm afraid I imagine you would you could see this as like even when we're doing our marketing on our side like we sometimes and we've not saying we sometimes like we've considered it is like intentionally putting errors in your in your response because you can tell when someone sends you an authentic email you're like okay this was like real short it was to the point and then they like versus you see this perfect email and you're like, okay, I think that might've been AI or it's really long and detailed. And so I definitely think there's gotta be that balance. especially when working with specific clients, they want that human to be on that one side. But there are other things that internally you can send to your team. Like, hey, send me an AI email. If my CEO sends me an email, it's AI, I'm okay with it, sir. I'm happy to say, just but to the customer, right, to the clients, that's, think, where the human side's gonna be, needs to be on the forefront. Speaker 2 (26:38) And you know where you have the opposite of that is I want to know who read my x-rays or my MRI scan. And you know what? I don't want a human reading it. I want AI reading it. I want some saying AI read your x-rays because it's way better at it than these x-ray techs out there. And so it cuts both ways. I want a human in certain instances, but in others where it's just Speaker 3 (26:40) Definitely. Speaker 2 (27:05) way more accurate, again, and scans it is. It's just that's real. You you talk about a full body scan and the whole thing's been, you know, the human looks through it, they can find stuff here and there and they don't know where to start or what. I mean, it's just too much data for one or even several humans to look at, you know, because you have you have to have a neurologist looking at the brain and the brainstem and you have to have, you know, you know, a different, you know, a what's a foot doctor called? Yeah, a psychiatrist looking at your feet, so on and so forth. Somebody else looking at your knees and looking at arthritis. Somebody else looking for cancer. So you'd have to have a whole team of people looking at an MRI scan to try to figure out what's going on and trying to get all those people to talk That's why you're built. Yeah. And then AI can just come back. Yeah. It's just say, it's probably cancer right there, which is consistent with this thing we saw up here in the brain. Speaker 1 (27:50) 45 Speaker 2 (28:00) You know, or there's some arthritis here and that's, you know, that's the way he's got arthritis is his wrist from hanging onto the mouse, know, 14 hours a day. Or do we all have our messed up wrist? Speaker 1 (28:11) Yeah, yeah. Before we wrap this up, and I appreciate you, Dan, Alfredo, for chiming in and sharing more perspective. For those who listening in, maybe want to learn a little bit more about Cinch and maybe more about what your team does, what would you suggest they do moving forward? Speaker 2 (28:27) Yeah, just jump on the website, cinch.me. We're always behind. We're trying to get as more information up there that we can. And just hit us up at hello at cinch.me, and we'll get back to you or check us out at the trade shows. But I think our new trade line. I'm going to try this out on you guys, right? So is it Sherwin-Williams? I shouldn't get this wrong. Is it Sherwin-Williams that says cover the world? It's in Mexico too. Speaker 1 (28:52) paint the paint the world or Speaker 2 (28:54) Paint the world. it says cover the world. All right. So I think for Cinch, because this asset management is so big, I think we're going to be like, know, tag the world. Tag the world. I'm talking about, you know, in hospitals, but we want to get to the point, Kyle, where we're handing these out so you can tag all your stuff internally. you know, I have my old skis sitting over here. And guess what? I have our little tags on it right here. So I know I can go look that up and see, these are my old monitors that I got back. you know, whenever I bought those in the 90s or I wasn't 90s, but in the teens and broke and all that. And so that people could just tag everything in their house, right? They know what they had. You get a TV, you tag it, you know, you have your wife is like, well, we're having a problem with the TV. What did we do last time? You're like, I don't know. called somebody and they fixed it. Like, what was the problem? I it was flickering and you know, I don't remember. Well, now you just scan your little tag and move on. So I think for us, that's the future is that everybody can keep track of every single thing you have, whether it's in your personal life, your business life, you know, and or whether it's medical equipment or construction equipment. So if you're trying to keep track of assets, send us an email and we'll help you out. Speaker 1 (30:08) Awesome. Great. Dan Alfredo's nice having you guys on the podcast today. And remember if you're a software development space or in that, in that company and your AI agents just hallucinating or just miss handling basic conversations, go to gpt trainer.com to learn more or how we've helped our clients fine tune the AI agents. Thanks guys for listening in. Dan Alfredo is a pleasure. I love having the conversation. Hope everyone has a wonderful rest of your day and we'll see you on the next episode of AI Chronicles. Speaker 2 (30:39) Thanks,
