Dealer Tech Tuesdays

AI, Data Security and Profit: The Dealer Playbook

• John Acosta

Share episode

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 1:06:37

AI, Data Security & Profit: The Dealer Playbook | Dealer Tech Tuesdays with Prasad Chavali 

In this episode of Dealer Tech Tuesdays, Prasad Chavali (CEO of Axel Automotive) sits down with host John Acosta (CEO of VTech Dealer IT) for a candid evaluation of data, AI, and cybersecurity in modern automotive retail. 

Prasad brings an advanced AI background dating back to 1995, now uniquely paired with VTech Dealer IT’s 14 years of custom dealership data infrastructure. Together, this partnership delivers a specialized technology built to help dealers win. In this conversation, John and Prasad unpack the exact blueprint required to transform raw data into a distinct competitive advantage and expose where most operators are inadvertently leaving revenue on the table.

📈 What you'll learn in this episode:

  • The 5% AI Rule: Why 95% of AI tools fail to generate lasting ROI, and what separates the systems that succeed.
  • The Data Foundation: Why unified, clean data is the only viable prerequisite for trustworthy AI.
  • The Security Horizon: Emerging cybersecurity risks in automotive retail and how to safeguard your operations today.
  • The Product Champion: What internal leadership looks like during a successful data implementation.
  • The Retention Engine: How customer lifetime value (LTV) data fundamentally transforms loyalty and profitability.
  • The Future of Operations: Why shifting from "running reports" to "conversing with your data" changes the game.


💡 About Our Guest: 
Prasad Chavali is the CEO and founder of Axel Automotive, a data intelligence platform engineered exclusively for automotive dealerships. Through Axel One, the platform consolidates data from hundreds of disparate sources into a single, unified interface delivering real-time operational clarity across sales, service, finance, and customer relations. 

🌐 Learn more: www.axelautomotive.com

🎙️ About Dealer Tech Tuesdays:
Hosted by John Acosta, CEO of VTech Dealer IT, Dealer Tech Tuesdays delivers practical, executive-level insights into automotive technology. Every episode features honest conversations with industry leaders about IT infrastructure, data strategy, and what it truly takes to scale a modern dealership operation. 
🌐 Learn more: www.vtechdealerit.com

 

Powered by Technology. Driven by People. Defined by Performance. 

#DealerTechTuesdays #AutomotiveIT #DealerData #Cybersecurity #AutomotiveRetail #DealershipAI #PredictiveAnalytics 

Support the show

Sponsor And Guest Welcome

SPEAKER_02

Dealer Tech Tuesdays is brought to you by VTech Dealer IT. Guaranteeing your dealership premier reliability, stability, and customer support. Transform your Tuesdays into a powerhouse of growth. Contact us at www.vtechdealer.com. Hi, my name is John Acosta. I'm the CEO of VTech Dealer IT. Today on the podcast, we have the CEO of Axel Automotive, Prasad Shivali. We're going to be talking about AI, the future of data analytics, and how you can keep more money in the bucket and keep profitability high for your dealership. Prasad, welcome, man.

SPEAKER_00

Thank you.

SPEAKER_02

I know we've been working on this for a little while to get this conversation going on kind of what's going on in the data world, what's going on in AI, what's going on in dealerships, and our partnership as an organization, you know how that how that's working is is is really exciting. And you know, I keep getting people asking, like, where did Axle Automotive come from? So where can you tell me a little bit about the story of where the company

How Axel Automotive Began

SPEAKER_02

originated from?

SPEAKER_00

Yeah, it's funny. You know, we started talking before AI came into the forefront. You're one of the OGs of AI. You know, like you studied AI and when we start talking about, hey, the data, how do we protect the data, yeah? It wasn't even in the conversation. Yeah. And uh anyway, yeah, so Axel Anomotive and it's funny. I was looking at uh on my original deck from 2019. We're gonna build a negotiating negotiating agent, you know, like online negotiator. Okay. Uh I still have a deck where during the nighttime, when somebody comes in and says, hey, I want to buy the car. So we started building this algorithm. If I want to get rid of this car or I want to sell it, how can I negotiate with some data parameters to kind of get them to commit to it? And that's what I pitched to somebody because we were working on this platform uh that has a lot of integration. So, like, how can we take advantage of the data? And then COVID came and there was no cars to negotiate. And we're like, okay, it's a cool name. We have the concept, we have all the data points, so what do we do? And uh so we kind of went back and said, you know, things are gonna come back when people need to start looking at the data to make decisions and how many cars they have. And we expected that to happen around end of 21. Okay. Uh so we started going back to the drawing board, and one of the guys said, Maybe we should look at service side. Okay. Instead of buying cars. So we we originally named our product uh uh service max or slipstream, okay, where we can look at all the data on the service side, yeah. The ROS, the customer, and all that stuff, and start building a max max the service revenue side. Then when we started adding that, I was like, okay, the inventory is coming back, maybe we need to have the sales side. And then the data started actually adding up. We had went after fixed stop fixed up data and then sales side. And then one of the guys said, We still don't have really good metrics on our financial performance. I was like, all right, let's add the financial data. Oh, wow. And then this whole CDP conversation is like, oh, do you have a CDP or do you have a customer data? And then we like look back and go, why don't we just get everything? So I then we did a uh a tech stack survey. I had my guys go through uh some of the newsletters and uh you know list every vendor and all that. Remember like even Steve Steve Greenfield puts that uh uh newsletter? Yeah. So we you know like 300, 400, between three to four hundred pieces of software. I said, so where is all this data? So I started talking to a customer and said, Do you know where all your sources are? And uh he said, I think about 17 different ones. And I said, and do you have access to all that? And he's like, Okay, let me see. And then that's that's when I said, Okay, we have a product where if we can bring all that together uh for the dealer, not for us, where they feel like I have my hands around it. I I thought there's a value. And some people said, No, nobody's gonna pay for their own data again. And I said, they're paying for their data multiple times to give it to vendors. Why wouldn't you pay it once to get it in one place and then not having to worry about where it is? So I it would it was a real tough sell in 2020 and 21. Because people were like still like, okay, we don't know what's gonna happen. Yeah. You know, are we gonna have enough church tough sell or are we gonna make chips again? You know, all these conversations were in place, and that actually gave us some time to build in uh framework together. And then the first customer that signed up uh was in 21 as a concept. We still had it bits and pieces, but it hasn't come together really together. And he said, Okay, I'm gonna be linked to. He said, How many days do you think we will see our first data? And I said, six weeks. And he said, We've been working on this for eight months. If you show me data in uh six weeks or eight weeks, I'll be your fan. Oh, jeez, that's amazing. Because they had their own uh internal data warehouse project.

SPEAKER_02

Like somebody doing our data like with Power BI or Power BI.

SPEAKER_00

Okay, exactly. Power BI, three developers, uh, all the APIs themselves, you know, put it in a cloud. But the API wasn't the problem. And once they have the data, they were all programmers, right? Then the CFO will go, Can I get my trial balances data? They're like, What's that? You need to tell us where that data is, so they will go. So he was really pressure that I'm paying this many developers, I'm getting the data, but it is not just in a usable format. So we showed the data in about four to six weeks time frame, and all of a sudden put the first five reports from the data, and at that point he said, Can I buy your company?

Nightly Reports Dealers Actually Need

SPEAKER_02

What were the first five data reports that they were starving to get?

SPEAKER_00

Nightly summary.

SPEAKER_02

Okay.

SPEAKER_00

Uh this was a group that had 20 stores, so every night every general manager were texting how many did that that day. Right?

SPEAKER_02

They have a group chat of text messages going back and forth, everybody saying the nightly summary. Exactly.

SPEAKER_00

And then uh there were somebody was putting it in a spreadsheet, but when you text it or when you email on the phone, you can open up the spreadsheet. So the owner is like, I need a group view and individual store view. So that was the first report we went after.

SPEAKER_02

So nightly summary. Group view and then individual store view. Okay. Because as an owner, I want to see my group, but or compare the forward stores to the forward stores and the toy stores to the Toyota stores.

SPEAKER_00

But we sent an individual sheet for each GM too. Oh, okay. So each GM is getting that individual sheet. Yeah. There was a combined sheet for the group. And as soon as that report came in, the next one is what are the three things that I'm looking after? Okay, what's my service contract penetration? Yeah. And he was selling another product called uh Magic Jack. I can't remember the product name. I said, I want to know how many of those we sold each day.

SPEAKER_02

Okay.

SPEAKER_00

So the report started going from, okay, here is your summary. Okay, now here is your contract penetration, here is your magic, you know, magic touch or whatever that product penetration. So that email started getting longer and longer, the nightly summary. And then the fixed guy came and said, How how come I can't get a summary? Nightly. Because they go back next day, look up all the postings. So the first five reports obviously ended up being the sales summary, yeah, survey summary, kind of like an inventory summary, you know, my traffic, you know, things like that. You know, basic things that everybody assumes that you know people should be. But it is amazing, even in 2026, how much of that data is still being aggregated manually for a lot from a lot of guys.

SPEAKER_02

And prone to errors and prone to compounding errors because they're putting on an Excel spreadsheet and they're you know messing up the formulas, and then it ends up being well.

SPEAKER_00

Because nobody's sitting in the night, you know, crunching all these numbers and go, uh, here is your report first thing. And there's always one guy didn't fill out his store, uh, you know, like 10, 10:30, he will call and say, Hey, how come you don't have your number? Oh, I'm I'm out selling cars. Yeah, I'm not at the desk. Then they had to find somebody else to kind of put the numbers together. So that's when we said, okay, Axle, you know, the one name didn't come through come until later, the Axle one as our product. You know, we were like thinking Axel Automotive, Axle Data, uh, Axel Mining, you know, Axel Extract. We were kind of trying to figure out what's our identity, what's our lane.

SPEAKER_02

Yeah.

SPEAKER_00

And then uh the name actually came from somebody else, one of my customers. So he you guys put all the data in one place, and I'm like, yeah, that's what we do. That makes all the sense. And it's like, so I can go to one place where I can see all the data, and then you know, we come back and said, I think we should call our product Axle 1. Yeah. Because that's one place that you're gonna go to get your data, your reports, and all that stuff, and that's that's the whole history of how the product evolved.

SPEAKER_02

You know, you you what you just said about, you know, I hey, I got too busy on updating my c my spreadsheets. Like, how can you manage an organization first on inconsistency, right? That the dealer principal has so many fires and they're playing whack-a-mole, right? And they're how do you how do you have the instrument cluster of flying the plane rather than doing an autopsy report that's 45 days later, right? That you know, um, and I I name him a lot, and he's part of our organization is Paul Jensen. Paul Jensen was they were managing 16 stores and they were doing spreadsheets, they were keying in spreadsheet for 16 stores, and then a dealer principal was 45 days later looking at data that was ancient, right? It's ancient history. So, how do you manage an organization with a 45-day lag time? And when I'm like looking at this stuff, and I don't want to make this an informal, right? I want to make this a real conversation about data. Yeah, like the days to today is let you we gotta see what's happen what happened yesterday, right? And what that cadence looks like.

SPEAKER_00

And then especially if we look at on a first of the month, or you know, what'd you do as of last night? Yeah. And people are still putting in the deals, you know, they might have been working until nine o'clock in the night. Yeah, some are still on the desk, and you get some kind of a estimate, and then they're all like, oh yeah, I put together 50 deals, you know, for the month. Because there's a big variance between 45 and 55. Yeah, there's a big variance. If you do pacing on every on everything. So what is the exact number? And also the other thing that uh very interesting this one is it actually exposed a few of the operational issues. Correct. Yeah. So what's the data is gonna do? All of a sudden it's like, hey, well, why am I saying only 47? I thought you said 50, and then where are those three? And we're already like, oh, where let's see where those three are. Yeah. Well, we sold it, the cars are the lot, but we nobody put it in on deal sheet. Yeah.

SPEAKER_02

Or it's like, oh, we had to unwind that deal, that's a next month deal, that's for this month, and it ends up getting fuzzy, funky mouth.

SPEAKER_00

So I will tell you, for the first two weeks we sent that nightly summer report, I would say half the people said that report is incorrect. Everybody's like, hey, my store's information is correct, incorrect. So the proof of burden was on us.

SPEAKER_02

Yeah, that makes sense.

SPEAKER_00

And so I had to go store by store. What do you think you're saying? Because data doesn't lie, right? And I said, I'm pulling from uh the actual uh DMS. Yeah. Uh and then we realized, okay, you're in Texas, you're in Vesco, so we need to start adjusting the timing. Oh yeah, we were here until nine o'clock, you know, last night. So we'll okay, we'll give you a nine o'clock the store time zone, yeah, and then run the report. And then it started, you know, cr making that difference smaller and smaller. And then one store is like, oh, we don't put in uh into the deal until we have a pay something like that. You know, we don't book the deal or it's delivered. Yeah, it's still impending. It's like how how will anybody know it is still impending? So amazingly, the owner sent a memo to all the GMs and said I want the nightly report to be accurate. I want you to close out to you by nine o'clock.

SPEAKER_02

Yeah.

SPEAKER_00

And if you didn't actually deliver the vehicle, it's okay put in the next day. But don't tell me like, hey, I sold 15. Well, two are gonna be picked up tomorrow. Are they today or are they tomorrow?

SPEAKER_02

Yeah.

SPEAKER_00

So that's when it's like the data is like the whole conversation is why is data important in one place? Whenever you're trying to look at the data and then say, I'm gonna make some decisions or I'm gonna change my flow, that's when it starts

Trust Issues And Implementation Champions

SPEAKER_00

coming handy.

SPEAKER_02

Yeah.

SPEAKER_00

You know, it's not just there to sit on or you know, just run some KPIs and you know put it on a dashboard, right? Which we do, but you know, where is it gonna start impacting day to day?

SPEAKER_02

The part that I'm that I'm that I'm really impressed by is the contextual data. It's like when you see the number show up in four or five different places, then you know the number's accurate or as close to accurate as possible, that you don't see these outliers that are happening. It's like the the the power of context of having all the information in one place is like that's I think that's the revolutionary part. That it's consistent, it's the one source of truth, and that you have the power of context. You have those three things that you can say, okay, now with confidence, I don't have to go into firefighting mode, I can go into managing mode because I have the data. And some of the best operators that I know manage through spreadsheets. They have their, but they have to go in and they have to key in all the spreadsheets. They don't they know their number to like ten doll off or to the dollar of what they're where they're gonna land at the end of the month, but they still have to do that process of doing that. Like, what's the resistance? What's the resistance that you guys are coming across that are saying those numbers aren't right? Or like you're getting arrows because the it exposes operational deficiencies in the organization. Like, what are those patterns that you see in the dealership a lot?

SPEAKER_00

The first thing, obviously, uh they will say the numbers are are they accurate, right? So we had to show uh that we'll be getting our are real or you know, like validated. In last year or year and a half, we had to spend a lot of time on that data validation team. Yeah. The point you made. Once we're confident, anybody can call us and say this report is incorrect, then we'll say, okay, let's go, because we have very 99%, nine 99.9% confidence that you know our data is validated to a certain extent. There's always some errors, you know, data file didn't load and things like that. But the confidence factor, initially, we used to be like on our heels saying that you know they're questioning this report. Yeah. Because since they know their business inside out, maybe they're accurate, they're correct.

SPEAKER_02

I think that assumption of they know their business inside out isn't correct. You know, it's like how could you possibly know your business inside out? Because everybody's protecting their things.

SPEAKER_00

Like you said, they've been using these spreadsheets for a while, right? Of course. So you are institutionalized to believe your spreadsheets. Yeah. And you're operating based on that, you're making decisions based on that. So the resistance is definitely the person or you know, a few people that are updating the spreadsheets are kind of scared to kind of change the process.

SPEAKER_02

Yeah.

SPEAKER_00

You know, it's like, hey, you're now telling me that I need to do something different. You know, like, am I gonna rely on your report to order the cars now? Um, but as soon as you find a really good champion, the data champion, it could be like a CFO, it could be a very progressive owner or somebody, and he goes, No, we're gonna trust in this number and let's kind of get to it. And then as soon as they push that, everything else starts falling in place.

SPEAKER_02

Yeah, it's like they the the train, you know, once it's on the track, starts getting exactly you know, getting straight.

SPEAKER_00

And that's one of the our challenges is to find that product champion. I know you guys uh do that for us for a few of the dealerships. Once you look at the value and go, I'm gonna be uh implementing Axel and everything else is just a um favorite game from that point out. It's like, okay, what uh what do you want to do now?

SPEAKER_02

So the the thing that I'm like that that I'm noticing is you know, no the best product in the world can't survive a bad implementation. Yeah. And so like how those implementations go in, because you see resistance across the board, and they're like protecting their little areas and they don't want to be exposed and they don't want the information to be out there. And so there's like um, I don't know how else to call it, but throwing sticks in the spokes, you know, of making those implementations happen. So having the champion that's gonna be like, no, we're doing this, we're burning the boats, we're doing this because we're doing this. And that that just creates all the momentum in the world. If you could, you know, if you could say, my ideal project champion, what would that person look like in the dealership?

SPEAKER_00

In fact, if this was uh part of our evolution initially, you know, there were multiple people that were the champions, nobody will take ownership, and then the two or three failures we had early in implementations were because of that.

SPEAKER_02

Yeah.

SPEAKER_00

And so now it's almost like we asked that it's the first question, who's gonna be my product champion? And uh in some cases, the deal principal said I would be the champion, and we were surprised, you know, because he believes in it, and there were a couple that actually do that still, you know, not huge groups, but you know, two to five stores and say, I'm the one requesting this product because I want to feel comfortable. And when you get those as product champion, his assistants and everybody will do you know all the all the work that we need to get done. Um so far I think a CFO or a dealer principal or like even a uh some of them have uh uh training people like our data people that you know took took the they know everything, you know, like who who to get access to. The latest information I'm doing. She's actually in marketing, but she worked with the GMs before. She knows who to contact to kind of get there. She's okay, you're in marketing, but you're helping us with Axle One, and it's working out really great because you know the person that is connected to and everybody's like, hey, why are you doing this? Well, the GM said to do this, or the dealer uh principal said you know we should do it. Yeah, and then uh everything falls in line. So that primary contact point that understands who signed the contacts vision is really the key. Yeah.

SPEAKER_02

And I I think on the on the that solid base of adoption and

Private LLMs Need Clean Data

SPEAKER_02

then of context with data that's already been organized, you can implement some really strong initiatives based on solid data, right? And I know you were just recently at a Sotocon and everybody's talking about AI, and we were talking, you know, I I know you bring come brought up something about uh cybersecurity that we're very much involved with and and having those conversations of what EI is leaking into into this world. What what are your what are your thoughts on on that world, especially you and tell me a little bit about your AI background because you're I know you're one of the the the originals on the on the AI side. So tell me about that and let's put it on the AI side.

SPEAKER_00

My AI pride was in 1995. Okay. Uh that was my master's thesis. You know, we had to submit a project, and they you know they said, you know, go do this right in AI. And I was like, you know, we all had a Terminator 2 or Terminator movie, and you know, Skynet was there. Skynet, right? Oh, is that what AI is? And then so I went and met this professor at Arizona State, and he was doing actually a project with FAA at that time.

SPEAKER_02

Okay.

SPEAKER_00

Uh to do predictive landing patterns. This was in 95.

SPEAKER_02

That makes sense. Because the aviation industry is a very heavy into automation. I mean, that's I took a class called the human factors in aviation, is literally if you automate too much, the pilot falls asleep and you fly into a mountain. And if you don't automate enough, the pilot falls asleep out of exhaustion and flies into a mountain.

SPEAKER_00

So so he gave me like a little bit of uh uh a small version of the project. So, you know, the programming language was uh Lisp LISP. Um it's almost like a blob. You start feeding the data and it starts expanding and start to re-analyze and then give you the output. Pretty much the same concept. Yeah. But the computing power was really expensive. So we had to preserve uh a cycle on the mainframe to run our program in the nighttime. Oh wow where it's not bringing everybody else down. And so I was able to get enough uh uh cycles to start feeding like samples of data and let it run through and speed out. And like in FAA, even at the time, if it takes too long to compute or analyze and give you the data, it gotta be almost instantaneous, right? You know. We're looking at twenty twenty six, the computing power is much more powerful. So we've been talking about AI since then, and what what really happened in the last maybe even two years, it became uh very uh economical, right? We can crash the phone. So before that, most of the AI was very good logic, very good algorithms uh that could run through. It wasn't learning as much as fast. So with that, when we had a pivot too. Last year we invested a lot into the AI side, but the easiest one is to kind of sign up with one API and you know, plug it into your data and start spitting out, you know, like your responses and say we have the AI. And uh I'm like, that's what is that extra value? What is that gonna give us? So I challenge our team to come up with like private LLM, self-contained, uh, only reading out our data. It doesn't need to know everything going on in the world. Yeah, uh it just needs to know what what's happening in that group.

SPEAKER_02

It's like intelligent enough AI, right? Exactly. It's like yeah, maybe don't tease it too much. Yeah, exactly.

SPEAKER_00

And then we decided we don't want to mix the data from groups either. Because what's it's learning in a group in the Northeast versus in California, it's not the same uh learning uh patterns. So we fear the data, but it's actually going to um come up with the different answers because it's actually taken into consideration uh the timings or brand, yeah, the value, the customer base, all that stuff. So it took us six months to kind of spit out the first really intelligent answers. And that's when the bulb uh you know went off and say, our AI is going to do things that we used to take a lot of reports to give the output. Like, for example, take that nightly summary, right? We sent this report. Uh, how many we did we do? All of a sudden, what are the other questions? How did we do last year? Yeah, you know, you know what so when you start asking the questions instead of having to go run the same report, change the timing, we can now use AI to give those all those answers. So we start creating this topic library. So you open up a nightly summer report, you can say, compare it to last month. Yeah, show me a trend on this. Is the new units the only thing that is uh you know going up? All these questions now could be used by you. That makes sense. So that's how our product evolves. So it's not a separate product. We don't rebrand our company as Axelon.ai. Yeah, yeah. Because it's a part of the product.

SPEAKER_02

But based on the true concept, and I think the real strategy is that clean data provides strong, strong AI.

SPEAKER_00

As what to Khan was uh somebody said the foundation for a really good AI product is going to be the underlying unified data. And then I was like clapping, you know. The only guy clapping on the that's exactly where we went. Uh we didn't have a fancy interface to present, but you know, the amount of effort we put in to kind of get to the underlying, and also he said normalized data. That's the other key. Normalized data, yes. Because if you start feeding a bunch of raw data, yeah, it's gonna take more and more uh uh time to learn. Yeah, and this whole world hallucination in AI space is real. Yeah, yeah, and uh even as you know, we're looking at it and say, and I said, okay, let's kind of make it really simple. Yeah, just start feeding very small chunks. Yeah, get rid of the hallucination, starting uh uh making it to respond. So if you ask how many new cars I sell last year, the same month, that doesn't need to feed 12 months of his in history if we can you know feed at a lower level. So when when we start adding uh components like this, now we rebuild a lot of these data models for the reports into manageable AI chunks. Yeah.

SPEAKER_02

So they're almost like agenc little pieces of a larger you said, hey, you're gonna be the fixed ops agentic little agent that goes on to it. So the data's got gonna be a lot more accurate based on your data set, not the world's data set and answering Then we start connecting those agents and say all of a sudden I switch a question, it knows where to go and get the um answers.

SPEAKER_00

So from that aspect, I looked at, you know, I went to the auto industry uh AI conference, the first of its kind, uh last week, and I was kind of curious to see what's coming out. And I mean, you know, the software is coming out fast. Oh, right? It's so dangerous. Uh it's like there are so many companies. I'm not, you know, saying any good or bad anything, but there are so many ideas, all of a sudden they're all coming to reality, right? Yeah.

SPEAKER_02

And then any any person with an internet connection can vibe code their way into uploading CRM data, and then suddenly you got a data breach.

SPEAKER_00

It's a funny thing. I saw a product and the guy next to me said, Oh, yeah, I can rebuild that in 10 days. And I looked at him and I said, What? He said, Yeah, you know, I can build the same thing in 10 days. And then that's when I started realizing what is going to be the differentiator. Because people are gonna start enabled to build if you give them all the components in 10 days. Yeah. You sign up with Cloud and all that stuff, you will you can build the application. In fact, there was a fun exercise during the first day. We all sitting down and you know, going through the exercise, and a solution, including the front end, everything was got got done in 30 minutes in that session. Yeah. And we're like, okay, so it's here, you're gonna have all this development. So what's gonna be? So somebody asked a question. So what's gonna be automated SaaS world's gonna look like in two years?

SPEAKER_02

Yeah, the SaaS apocalypse, right?

SPEAKER_00

And then somebody said, why two years? What about one year? Yeah, it's like, you know, how are we gonna wait two years? And then everybody's like, we can't tell. But it may not be exactly SaaS model, it's going to end up to use the AI, right?

SPEAKER_02

Yeah, it's gonna be like AI as a service type type of situation where you have chunks, pre-built-out chunks that you put in. The the piece that I've been thinking about a lot lately is like, you know, we come as as IT guys, right? It's like you walk into a dealership and some programmer or some guy that was an IT guy that was a service manager wrote his own code about gas tickets, and then he left, and he's like, our gas ticket systems collapse, and you know, they're that's gonna happen times 100 with whoever's creating these micro little bots that are gonna morph into these unruly monsters that people are gonna, you know, the car industry has 40-60 percent uh turnover rate. So the people are gonna leave with these pieces, and your data are you don't you're not gonna know what's connected to your database? Like it's I think in a pending disaster that's gonna come.

SPEAKER_00

Yes, you know, but I don't want to look at it more from that cobbler, you know, that that'd be apocalyptic, yeah. Apocalyptic view. Couldn't say fast enough, but the responsibility is going to be uh what needs to happen now because one of the guys in that conference said in the next couple of quarters there's gonna be a huge data breach. Not necessarily in automotive industry, it could be anywhere.

SPEAKER_01

Yeah.

SPEAKER_00

And somebody somewhere will say, Oh, we didn't know we fed the data to the AI. And even with the best of intentions, it is growing so fast. Last week's you know, models are not good this month, and who is who's wanting, you know, they want to win the race to the finish line, so they're gonna be anxious to put out

Cybersecurity Guardrails For AI Access

SPEAKER_00

something to say, hey, we got this product. And he said that's going to create a little bit of reset for us. Yeah. And how the data is going to be shared. Yeah. And that's where we feel really good. We're going to bring it back to reality and say one place and one AI engine. Yeah. You don't need 20 products using 20 different LLMs, and they're all analyzing using their own algorithms and coming back with that data.

SPEAKER_02

And and I think the important part is that it's agentic. So what that means for people that don't know what we're talking about on the A side is that it's just a small AI engine that is cut off from the internet and it's only on your server on your data. So it's like, you know, you have this little nuclear reactor for your data, right? It's not like you're gonna have a massive radioactive spill out there because you're connected and somebody can can create a problem.

SPEAKER_00

In fact, you want to create those small agents because you want to contain the data. You don't want it to read everything. You know, financial shit data shouldn't be in any of these uh containers. And then to add to that, remember when we set up the Axe Mon data, we were using the users' roles, what do they have access to? We need to apply the same exact to the AI too. Yeah, that makes sense. Just because the data is there, you cannot say, hey, how much is my boss making?

SPEAKER_02

Yeah, exactly. It's like what's my the pay plan for my general manager over the last two years?

SPEAKER_00

Of course. So the guardrails had to be built at the user level, the department level, the dealership level, at the group level.

SPEAKER_02

Yeah.

SPEAKER_00

Right from the beginning. Yeah. Yes, I can put up a much uh fast product in AI tomorrow, but again, I had to go back and you know look through all these uh uh little nuances because we put that together. And remember, that's how that's where you and I started talking about. Yeah, and I said, we have the responsibility to protect all this data. Correct. And I don't think we could do this by ourselves. So we need to start partnering with people that are in the cybersecurity infrastructure, uh, the MSP services, like the turnover, who has access to remember like we use the Windows login, so when somebody leaves, the access gets cut off. Yeah. And uh I think that that is still true, and it's gonna be more important in the future.

SPEAKER_02

Oh, absolutely. It's the human is always the vulnerability, right? It's like the human and the management of stuff that's and what's even more dangerous in this world is that it becomes everything becomes an assumption. It's like, oh yeah, that's protected, and it's like out of sight, out of mind. And I think you're in the that next six months that something's gonna happen because you know about our partnership with uh with the guys from Dymum that they come from the data gate, like the AI gateway security from the intelligence community. It's a it's a big deal. And man, they're saying that there's leaks everywhere.

SPEAKER_00

It I still get some pushback on how we set up our product, access into the firewall. Like, you know, I come to you and say, what is the dealership network? You know, we only allow uh access to based on where they're at. There's still so much so much software out there that you can log in using your simple username and password, no restrictions that has customer data, you know, financial information data.

SPEAKER_02

Yeah.

SPEAKER_00

The guy said, Oh, I'm traveling, I'm in a hotel room, I want to access Axle One.

SPEAKER_02

I'm like, Yeah, no. Don't ask me, ask your ID guy.

SPEAKER_00

In fact, we kind of you know put the responsibility back on the security policies of the dealership and said, if your IT guy signs off, that my so-and-so can have access from anywhere, anywhere in the world without a VPN and a MFA, then it's not my responsibility. It could be their responsibility.

SPEAKER_02

But it's like just sign right here. Yeah. You know, releasing me of all liability.

SPEAKER_00

And uh so right from the beginning, we started using their security for login, MFA, whatever they use, and their VPN and their security policies, retention policy, who has access and all that. You know, people are like, why do we need all this? And it's like, that's that's where a solid implementation needs to look at all these aspects. Uh if I hesitate and run through the implementation quickly and don't do all the work in the long run, it's gonna come back and bite us somewhere. Yeah, that we didn't look at this one. We know we are we now during discovery ask all these questions.

SPEAKER_02

Yeah.

unknown

Yeah.

SPEAKER_00

And so far nobody said, I want to open up this to everybody.

SPEAKER_02

Yeah. Once they understand the liability of what it looks like, yeah.

SPEAKER_00

And you know, if there if it wasn't an option, you know, where they said, hey, you can only access to a URL and you know, I don't have another option, maybe you'll think and and like you said, the uh the data preach and all that imminent threat is in the back of a lot of people's minds now than ever before. So they're they're also asking the questions like, you know, how how is this uh secure? Yeah, there's one guy that said we don't want to really take the responsibility having the data in our own premises because it opens up a lot of liability and all that. Yeah. And so we had to show them the measures that they're not going to, in fact, it's more secure. They will have a copy of their database. Anything happens outside their environment, they're not going to be out of business. You know, uh we never go in and say, try to sell the fear factor. Yeah, yeah, but we are selling the comfort factor, knowing that if anything happens, there is something here.

SPEAKER_02

Yeah, it's a peace of mind aspect of it. I mean, there was a lot of dealers that were with a large DMS data breach that happened, you know, a couple of years ago. Everybody was out of business, and there was some people that didn't have access to their data that they were like literally stopped for three weeks, you know, and and it that was that was really rough.

SPEAKER_00

We saw a lot of piece of paper. We went back to the 1960s.

SPEAKER_02

Yeah.

SPEAKER_00

Let's yeah, let's talk about deal sheet. Oh, that's a real critical deal.

SPEAKER_02

Jeez, yeah. You know, and and and some people were like, I don't know how to write a deal. Like, I have no idea. It's like, yeah, you can you can do this, the the and like writing up repair orders on uh by hand and doing all that stuff. Like people, we had some groups that stopped altogether and were like, hey, hey, we can give you kind of these sheets and do the stuff, do it this way. And there are other ones that were like, Johnny on the spot, they just they they didn't miss a beat, they didn't lose productivity in sales or in service. It's like it just depends on the operational maturity of the organization.

SPEAKER_00

You know, but that also added that little uh um need for issuance, yeah. Yeah.

SPEAKER_02

Well, here's the thing too, it's like you can't have your cake and eat it too, is you can't have full control to say if I want to take my ball home and I want to go home and I want to go with somebody else and not have the liability as well. Like if you're if you're gonna secure your environment, that's that. But if you're gonna say, Oh, I'm gonna go with another provider where you put it in their data center or their cloud or their whatever, they have control of it. And in the fine print of that paperwork somewhere, it's gonna say, Well, you gotta, you know, trickery of what's what that's gonna look like, right? So that could potentially 100% be the case. But if you're you own your infrastructure and you said this is yours, if you want to grab this and you want to get your own development team and then build that stuff out, then go for it. You already have the data, it's already organized, it's already been built out for you. I don't see why somebody would do that because the product's great, but you know, that you have at least that comfort and that peace of mind that that can be a possibility.

SPEAKER_00

The last thing that I was gonna finish up in that whole SAS, you know, what the software with the AI is going to. Even the dealers are like, am I gonna pay for this, you know, per user license fee, or can I just build it with the AI myself? It's almost the question comes up, right? So you will build a guy, you know, you pull a guy from the backup and say, Hey, this one is a fifteen hundred dollars a month, can you build it? And they're all gonna say, Oh yeah, I can build that for you, right? So it's it's tempting, but there's also those inherent risk of what are they signing up, what put in you know, what uh LLM or what the model they're using, where they're coploating the data.

SPEAKER_02

And it's worse because if that person leaves, you're grabbing something that's half baked or 40% baked, and you like suddenly a critical part of your organization is built on this wobbly bridge that some you know 22-year-old kid was vibe coding their way through. That that's a recipe for disaster.

SPEAKER_00

So I was thinking we should build a base for these dealers. It is your data layer, it is your infrastructure, it is your security layer. Yeah, use this to build your solutions. Correct. Exactly. Don't start from scratch. Don't go outside and sign up with something. I think that's where we we can really take this to the next level after that conference I was gonna talk to you about. Is I'm sure there is somebody that says, hey, I this is my data, I'm gonna build some you know, utilities on top of it. Yeah, yeah, yeah. But if we can give them this tool set that is, you know, uh vector.

SPEAKER_02

There's uh large language models don't have access to an MPI. All the PI is protected, redacted, obfuscated, whatever that is, like absolutely.

SPEAKER_00

And then if they use that N1 to build in that in that private, you're much more uh secure and comfortable, you should be comfortable.

SPEAKER_01

Yeah.

SPEAKER_00

Um that's where I feel like uh in a year from now, maybe six, it could be even weeks, there is gonna be a consolidation. Yeah. Uh almost everybody, a lot of uh I seen a lot of uh responders, communications, you know, video, all these technologies changing you know rapidly, right? Who's gonna survive? Uh, who's gonna have an advantage, unfair advantage? Yeah, yeah. Right? Uh and who's gonna be able to charge enough. Like, is it gonna be ten dollars enough? And then the other one is even the computing power is um cheaper now. Yeah. If you sign up with a paid model, an APA model, it's gonna start adding up the bill too. So that's one of the reasons the private LMs are gonna be affordable in the long run. Uh there are a lot of users. Yeah. And uh if you just want to give every person in your dealership access to the AI, you know, everybody says, Hey, how are AI gonna help you know, help my job? Yeah, you know, we definitely need to provide this to the our customer base and say, we're giving you the infrastructure, we're giving you the private, you know, agents and LLMs, we're giving you data. Yeah. So start demanding that your vendor is gonna use this to reduce your cost.

SPEAKER_02

So I'm I'm kind of I have half a mind. That's like, you know, that there was that uh MIT article that showed up talking about 95% of II companies, and it was like last year, right? That per failed to provide value or ROI because they're dealing with the front end of the house, right? It's like they're creating all these new visualizations and stuff. It's like that's like a we're looking my my suspicion is that we're looking at the problem the wrong way, right? Is that the value is in the 5% that you can say I don't necessarily care what happens on the back end as long as I get served up the best

Customer Value Insights That Drive Retention

SPEAKER_02

data at the right time, right? So, you know, I I've always had this example like a BDC agent, right? It's like you want the BDC agent to be replaced, or do you want them to go, or do you want the you know, the AI side to go through the 3,000 leads that they have and serve you up the 10, 15 people that you're gonna create a human connection with, that you can provide a lot of value to, that you can put them in the right car, that you're gonna create a customer for life, that you're gonna build a lasting relationship with them and be their go-to persons. Like I wanna, I want that to be, I want AI to be empowering all the mundane stuff that it's great at pattern recognition, it's looking at that side of the business rather than replacing the person that's in front of you.

SPEAKER_00

That that article is exactly what I kept saying. You know, maybe initially, let's say you implement an AI solution, they're gonna say 25% increase in response rates. You know, whatever those metrics are. Yeah, maybe first month, maybe second month. Is that gonna be 25% every month going forward? There is not enough uh efficiency to be gained. Yeah. So that's where I you know yes, our calls increased, response rate increased, we're making another 5% more on the service drive. Um those areas are still there to be you know to gain, but that's not a product that's gonna do that, you know, day in and day out. Yeah. Yeah. And the 95%, some of them may actually give you some uh instant improvement. Yeah. Right? But it's tough to sustain. And I cannot charge you a big monthly fee showing that one month, two months, or maybe three months, right? You you and I know like the market is saturated enough that a store that is selling 200 cars is not gonna get to 300. Yeah. Unless it's some major revolution, right? Because the market share, the competition is alright. Even in service, you can make more money, you can do mobile servicing, you can sell parts, but you're not going to grow a significant percentage, right? So when you buy the software, when you implement AI, the cost of software should eventually be showing up in the long run too. Yeah. Yeah. And where that efficiency is going to show up and say, it's not going to replace the humans, but your example of BDC, what are you trying to gain from that? Exactly.

SPEAKER_02

And there's something to be said, and I was having this conversation on my morning clubhouse rooms. It's like, what for your PMA is sustainable? You know, like in your PMA, you have this market rate. It's like this, this at the humming, you know, like the Goldilocks zone of this is going to be a 270 store. And so maximize 270, right? Maximize the grosses, maximize the back end, maximize your your customer retention, maximize the customer journey of the experience, maximize the herd that you're getting to. It's like we're going to stay at that 270 and make this um uh a machine that can hum along in a 270. Maybe you're in a mega market that you can be, you know, there's some stores down here in South Florida that are twelve hundred, right? That's they're monsters, right? Those are those are those are crazy operations. But what is that because there's nine million people in South Florida, right? But if you're in a small town, you're gonna be like, this is this is the thing, and we're gonna make great margins at this, and it's gonna be an efficient and an effective operation.

SPEAKER_00

You that word efficiency is what the whole technology is gonna give you. Yeah. The other one I heard about is the customer experience. Like I can now tell you if an email or response is AI or somebody wrote it, right? Even if there is little variations and all that. People are gonna get used to it, right? You know, if I get an immediate response, like, oh, did somebody actually look at my message between the two second time frame I spent and wrote this long email? Yeah. So still the human element is gonna be important. And I saw Dukan was also the human element, you know, the year of the human. And when you can implement that, a better experience, the loyalty factor, when they're gonna come back and start analyzing the data more and say, John's been with me 11 years, he consistently bought the cars, you know, let me take care of him, you know, a little bit better and you know, make that connection with you. And we started analyzing the commercial customer data for one of the groups I worked with. All of a sudden there were a few names that missed because they went to different stores, then we put it together, and they're like, these guys are top five. Oh wow. Not like top ten in three or four stores. If you add those three or four stores because they were purchasing and in a different phone number, nobody can put that together. That's where we put the AI data and say, This gotta be, you know, your priority to take care of them. Yeah. And then the other question we use the AI to say, give me the top five customers for my group or my store. If you go and ask me and say, hey, do you you know, do you know your top ten customers? The GM will go, yeah, I can give you one or two names. It's it's not their fault, but they never have access to the data in a way to say, hey, this is a good customer, this is a valuable customer.

SPEAKER_02

The customer center insights are next level. I mean, that is to be able to have answer for people I don't know on the podcast, right? That are listening is Axel has a customer insight or a customer center, right? That has lifetime value of the customer, how many times they showed up in service, how many times they showed up in sales, what they responded to, what their lifetime value score rating is over. It's like, okay, give me the customers that are a lifetime value of 80 or above that have never bought tires from us, right? It's like, are you gonna market and send mailers to the customers that are a five lifetime value score or a 10 lifetime value score that are gonna, when they get those mailers, they're gonna go like this into the trash, right? They're gonna throw those things away. You're literally wasting money. It's like, how do we better deploy capital to acquire the customers or take a customer that's a 60 to a 65 and bring them to a 70 to a 75? The value proposition in that stuff, and that the that you can actually quantify the health of the herd really gets exciting with kind of segmenting that data and how you guys are looking at it.

SPEAKER_00

And I I admit that we probably hasn't uh exploited all the data in the customer center to the fullest extent yet, because there's not a lot, you know, there's CDPs, there's a lot of marketing agency, there's a lot of activity going on in there. But recent conversations, some of them we had, very exciting in in that aspect. Imagine the scenario, you get a hundred people to the service drive. Out of the hundred, how many are your V VIP customers that came in today? Yeah, yeah, exactly. Okay, maybe you can run a report, but how would you embed in the process the service advisor gets a screen or report saying that hey, when these five people show up, go get the GM.

SPEAKER_02

Yeah.

SPEAKER_00

Have them come and just say, Hey, good to see you again, how is this doing? And then give him the data saying that I I know you still have your two or three cars. Give them all the data on the screen, you know, on the screen to you know, talk more sensibly.

SPEAKER_02

And I was talking to Brad Wise about this at Fermin, and I was like, what if you could say today your your service is complimentary? You know, customers spent $600,000 with the organization. Like you we're gonna spend $400 on. Yeah, we we took over Yeah, exactly. We took every we're gonna upgrade you to the most luxury truck that you have because you have this super valuable customer, and that glaring at this customer value is a 95. They service with you, they buy with you, they traded vehicles with you, they gave you reviews, they give you the contextual information of everything that you're doing.

SPEAKER_00

But see, in our right now, we have this loyalty and point system in everywhere, right? Airlines, hotels, even sandwich shops. Yeah right. A lot of the dealership tried it, you know, like rewards program or loyalty program. Now I haven't seen one that is groundbreaking yet because the complexity, the touch points are slower, you know, you come in for service maybe two times a year, so it's not volume-based. So they attempted and they left it in, they left it out. I think it still needs to be, like you said, the V V IP. Yeah. Add to Apple Wallet, you know, John is my VV and I come to the dealership, I scan it, and all of a sudden the lights go off, and you know, bullet. Yeah. It's gonna make you feel special. And even the most, you know, the customer that can afford all these freebies, it's not about the freebie, right? It's not the free sandwich like the Lenin, you know, in Seinfeld. It's the it's the feeling, right? The personalized touch. Personalized touch. Yeah. Hey, that uh the uh uh rental car is on us because you're a value. I'm like, oh good. They know that I'm a valuable customer.

SPEAKER_02

Or or you get a gift card for $500 for the Capitol Grill and said, well, like to take you and Mrs. Whitmore to, you know, to dinner or Mrs. Jones to dinner, and you guys get uh like we'd love to comp you know give that on on us.

SPEAKER_00

One of my first uh customers that I was talking to that uh came up early in the conversation, they now do this for their top they call it I think the top hundred across the entire 40 groups. Okay. They even have them like a luxury experience that they can come and stay two days or three days on us. They send this email saying that, hey, come to Vegas on us, go stay at the house. Oh, jeez. You know, they look they calculate the value to say, and you know, when those guys call, they don't even ask uh any questions.

SPEAKER_02

Yeah, but concierge like a concierge desk.

SPEAKER_00

Yeah, bring bring in the car whenever you're comfortable, Mr. You know, Mr. Customer.

SPEAKER_02

We'll go pick it up for you. Don't even worry about it. Like when it's comfortable for you, we'll go do the white glove service for you.

SPEAKER_00

And as the technology uh improves, uh the online you know competition comes to the mobile service comes in, these are the small things that are gonna help you retain that customer. You mentioned that 270, you know, per store. Even you can stay at 270, but most of your customers are your high-value paying customers. Yes. Doesn't cost you a lot more to acquire new customers, retention. Um that's where I think we need to, you know, help the guys with our data to say we may not help you to serve another sell another 100 cars, but we'll help you to kind of increase your profit margin. We will help you with retention.

SPEAKER_02

And and Brad Brad Wise talks about this. Not how much money you make, it's how much money you keep in the bucket.

SPEAKER_00

It's so funny you mentioned about the tires. I asked the dealer, uh, one of the dealers said, why why tires, right? You know, you there's so many tired places. Selling your tires from the dealership is a sign of loyalty. So they're actually trusting you. Yep. So you really don't need to make or mark up, you know, because that says, hey, I'm not gonna take care of everything. You don't need to leave here, try on the street, and ask another guy and compare shop and all that. At the end of the day, it may be a dollar or two difference, but look at the confidence that it comes, you know.

SPEAKER_02

They sell at a cost. Like for Atari's like, we'll sell at a cost, but they upsell and the relationship, and that's really because that you build the confidence to say, you know, my my mom's a very, very particular buyer, right? She wants to deal with Jim at the Honda place, right? She has a Honda CRV, brand new Honda CRV that she brought from Mergado store, right? And she's like, I want to deal with Jim. And that's the only guy that and I'm like, no, he takes care of me when I go there, but she has something that's that is worth

Expense Control And Marketing ROI Visibility

SPEAKER_02

a billion dollars. He has her trust, right? That he's gonna take care of her, and he's gonna put the winter tires. She lives in Chicago, right? So she's gonna he's gonna put on the winter tires or he's gonna take care of her, and if this stuff, he that trust is all the value.

SPEAKER_00

We do that in other places, you know. The cell phone company we stick to, an airline company, a hotel brand we stick to. Yeah. As soon as you feel that trust factor, yeah. So buying tires and all that stuff. I think all this noise about you know things are cheaper, the dealers actually charge you more, all that happened maybe a long time ago. Yeah. Nowadays, with everything transparent, I'm still a brand loyalist. Yeah, I will go back to that dealership or only to the brand one just because I know the fact that uh they will take care of the whole thing. Yeah. And we we all know everybody, even at like independent service shops and all those guys, you know, can do a good job. But if it's a new car, like if it's an asset, nowadays what car is what $50,000 to $100,000? Yeah.

SPEAKER_02

It's a big it's a huge amount of it's the second largest purchase that you'll make in your life. It's massive. Yeah. You know? It's it's absolutely massive. So what are you seeing of of what the next evolution next evolution? What do you think that this is gonna go to of what you're you're seeing in the market?

SPEAKER_00

I actually feel really good about where we're at. I think the evolution is gonna happen so fast. Uh all the things that we never thought are possible are gonna get you know reality soon. Oh, I never had to be able to do that data, I never able to do this. All those things are gonna happen. We're at the forefront of having the data on the right technology. Um we need to start watching what's gonna happen in the industry. Like in 2020, it was about no cars, and then we went through multiple phases, tariffs, you know, extra. We have to be in a place where we can react to that. Yeah. And I think we feel like we can react to that. What is everybody's gonna look at in the next, you know? I I said this in fixed up uh fixed up's uh interview. Expense control. Not necessarily go and slash and you know, slash everything. But if you really, really look at small minute uh expense cards, that's where your margin is gonna be. Yeah, yeah. One of the guys said my uniform uh uniform guy charges me, I don't even know. It's almost like I had to keep watching on it. But if we started using the technology to put those guardrails where they don't need to think about, hey, I need to go and check what my expense is. We're going to proactively analyze and say, hey, there's five things that you need to be looking at. Well, here's the biggest thing advertising.

SPEAKER_02

It's one of the it's by far the biggest expense in the organization, and you don't even know if it's working. So it's like you sold 270 cars. Why did you sell 270 cars? Yeah like that's those are the dangerous questions that people like get real. It's like, well, everybody attributes a little bit to that, you know, to that lead source. It's like, I don't know, it came up from a little bit from here and a little bit from there, and what it touched everything. It's like, if you can't see, and and the part that really kind of blew my mind is opening the Google Analytics side and then putting that in the same screen. It's like, well, now I can see my advertising, I can see how much we're spending for that lead, and what the source channel I'm I can't, I'm not a marketing guy, but like see that source channel is like we're spending all $25,000, $30,000, $50,000 on these channels, and this is what the you know, the how they're landing on our page. It's just the ROI is not there, and there's so much money that's lost and leaked in the organization because of that. Yeah, that contextual data is what's really powerful.

SPEAKER_00

The problem there was uh there were so many players it was tough to analyze the data. Yeah, that it was no detail. If you if you say I spent $25,000 on uh Google analysis, you know, on Google unless it was completely tracked and tagged and all that stuff. So there's still a lot of gay area. You're right. That is another area that we can you know show more and more data analysis to for them to make a decision. I'm one of those guys that will never come and say this technology is bad or the technology is good. You know, it's like we don't want to push it. Let the information be the source that they make their decisions.

SPEAKER_02

Yeah.

SPEAKER_00

I won't go and say don't you know spend money with that guy or you know, spend here. Yeah. You if you look at your data and then you go, this is converting at this percentage, and all of a sudden you can make the decision yourself. Exactly. It's just the power of context. That's that's what we want to do with the data using AI to say show activation of the data. You don't have to run the report. You know, one of my friends, Joe Shaker, said your tagline is stop running reports, start talking to your data. Yeah. And I think that's where we're headed. Yeah. If your data principle, if you ask the data, what is the next two or three areas that I need to start focusing? Yeah. And we analyze the data and say, you know, based on the trend, you know, these lines in the income statement seems to be a little if you start giving those kind of uh uh analyses right away, real time, yeah, using AI technology, yeah, they're gonna be sitting there and talking to it. Yeah.

SPEAKER_02

Well, this is the piece that blew my mind the most. Is so we're we're partners with GoToConnect, right, on the phone system. It's like it's not doing AI receptionist, it's not doing AI on the you know, person lit picking up the phone, it's doing the sentiment analytics. So you're saying when little little bubbles will pop up of like, you know, you had 3,000 calls this month, and the bubbles are some are green, some are red, and the red one will say wheel. Like we have a wheel problem, right? Talk to your manager, talk to your sales manager and say, we got the wheel problem, and we've had 400 calls of somebody sent getting really pissed off about the dealership about a wheel problem. The the over mental real estate and the overhead that it took to look at all those calls and to do figure out, you know, you'd have to say to your general your um sales manager, say, Have you listened to all the calls today? It's like, no, no, no, no. You don't want them to be listening to the calls. You want them to have a dashboard and you want them to look at the heat cases and the biggest bubbles of sentiment to say, we we have a wheel problem, or we're good at setting appointments and scheduling appointments, loaner car may be red, and like, hey, we got a loaner car problem. We may have a problem with scheduling, our online scheduler might be broken and getting the the data points so you can do and manage and be as effective as possible. And I've always had a problem with the word efficiency. Like the word efficiency kind of has like doing more with less. Like, I want to be effective, I want to make be efficient in movement, you know, like a swimmer to say, I want to make a I want to make the most outcome with the minimum amount of movement.

SPEAKER_00

I I'm completely in uh alignment with that. But the only slight difference there is if you say that's

Sentiment Analytics Without Alert Fatigue

SPEAKER_00

bad, it's red, how are you going to help them to fix it? Yes. Right? And there's a lot of tools out there that is already doing the sentiment red and all that stuff. Yeah. And it's gonna get overwhelming. I'm gonna see a bunch of red bubbles when I come in the morning. I have a wheel problem, I have a point. But that's where I think we need to connect the data to show the trend or areas and give it a little bit more solid you know, reference or context of what you use multiple times to say, yes, you may have a Wii O problem, but is it related to a single person? Let's look at the calls. Yes, you know, it could be one guy in the whole mix could be causing those red bubbles. It's not a problem across the dealership that we need to you know pull the pyro alarm saying that we have a problem. And that is where again connecting the data to it is the first part is already now done um tremendously efficient already. I see in the product sphere, the sentiments and analyzing call logs and all that's happening all there. So that's when I asked one of the guys, what are you gonna do next? Yeah, you know, start don't start throwing a bunch of dashboards with like numbers saying that you had you know 7% uh call with red. Great, all that is going to be that that analysis of data is gonna happen so fast in the coming weeks. You need a structure for the guys to start addressing one by one based on the impact to the business, too. Yeah. Sometimes you know there could be a lot of noise, but it's not really a problem. Yeah.

SPEAKER_02

Everything if everything's a problem, then nothing's a problem, right? It's like, oh, everything's red, everything's on fire. And and and that's the the part that I like about the context. Like if you see it in a couple different places, if you see it in your CSI scores, if you see it in your phone system, if you see it in your financial data, if you see it in Axel, you see it in everywhere, you're like, okay, there's a problem here. There that this is gonna, and what I what I'd love to see expert do in the future is to be like, we've seen this in a couple places, like top priority issue, and not create 700 alerts, just create one or two or three alerts because there's alert fatigue.

SPEAKER_00

Yeah, ranking them by priority. Exactly. A human being is only gonna be an ally five, maybe ten.

SPEAKER_02

And and to say, like, if this is where your biggest opportunity is to deal with this and it's gonna move the needle the most. Exactly.

SPEAKER_00

That's the word I was using for biggest opportunity.

SPEAKER_02

Biggest opportunity, yeah. Yeah.

SPEAKER_00

No, so so to bring it together, I think where it's gonna happen, you know, like you asked where we see this happening, we're gonna have a lot of opportunities to help with our data. Yeah. Uh even we had to look at the biggest opportunities. Where where do we want to provide some help right away? Yeah. Rather than trying to build all these products, you know, some could be like just bells and whistles, but it may not move the needle. We're gonna start focusing on what's moving the needle. Yeah. And uh I know we want to be very agile. Yeah and our products now gonna evolve in a way. In in the Discord, it's so funny. You know, it's like what is your pain points when you start asking that? That itself gives you the playbook, right? Of course. They will tell you, yeah, right. You know, yeah. You know, and uh uh it's fun. You know, I I'm really enjoying that aspect of it. Looking at the problem-solving ability using the data, and if we can stay in that lane, I mean I don't think you will ever run out of problems. It's not gonna be like, hey, I have zero problems. Yeah, yeah.

SPEAKER_02

As long as there are people, there's systems, we can solve it. There's always gonna be something. Yeah, there's always gonna be something. So, Prasad, what is a question that you think that I should have asked you or something that a cut topic that you want to cover?

SPEAKER_00

I think you covered a lot. I mean, we talked about the data vulnerability, the security, what's coming in, you know, what's coming in the future. Um, I think one thing that we should uh ask every one of our common customers or prospects is to start doing an audit or set up a AI policy. Yeah. Uh as much as it sounds like a compliance kind of a document thing, I think uh just to even ask the question. Uh there are a few organizations doing that already. You know, where are you at? Where you should we go and give them a blueprint to kind of follow. Because if you start embracing it too fast and then you just don't follow the steps, you know, we don't want to regret. So you guys can really help with that audit aspect. Yeah.

SPEAKER_02

Yeah, red teaming AI, right?

SPEAKER_00

Yeah. Yeah. I guess interesting. So I think you know, we should talk to everybody and say not not from a skating perspective, again. No, from uh you want to cover, you know, cover all the bases. So that that is something that we should be putting in front and center in in uh in our conversations.

SPEAKER_02

Yeah. And from a place of responsibility and from a place of creating momentum and alignment in the organization and all the organizations that we serve, I think that's a huge part of it.

SPEAKER_00

Yeah, if I ask that question, it sounds like, hey, here is a vendor that is asking the question. Whereas if it is actually part of like an your your MSPR, you're in charge of

AI Policy Audits And Closing

SPEAKER_00

if you say, hey, we're already helping you to build a um solid platform with no vulnerabilities, I think that'll be a huge impact. And we we need to we need to push that. Yeah. We do.

SPEAKER_02

We do. Prasad, thank you, man. This has been outstanding. I know it's been fun.

SPEAKER_00

You know, our paths crashed, you know, in a way that it's mutually beneficial, right? I never want to be in your space. I don't know any of the MSV, you know. Yeah. The more I can take my mind off the security part and you know, the technology, how fast is involved, evolving the source and all that. I really want to focus on that side. You know, I'm glad you're in your space and I'm on my space, and I think this is a huge uh uh win-win for both of us.

SPEAKER_02

I love it, man. And I'm I've always been a fan of data, of free economics, you know, like the free economics in the data, what is it gonna look like? And I I just think you guys are doing something really cool, man. We've been talking about for years, like money balling your dealership, man. What does that look like?

SPEAKER_00

It's almost like one guy said, You tell me you already have all this. Like, yes, it is a reality.

SPEAKER_02

So it is a reality. That's cool. How can people, if people want to get a hold of you, how do they do that?

SPEAKER_00

You know, nowadays everybody's on the net, right? You know, axelautomotive.com, but referrals is the best thing. And somebody asked me VTEC and Axel Automotive uh logo at NEDA, and I said uh you can find us through VTech's website or vice versa. Yeah, and we're gonna work with other people, but Axelautomotive.com um is the place to find and also LinkedIn, right? Nowadays we're all connected there and posting and sharing our thoughts. Yeah. Well, thank you for coming, man. I really appreciate it. Thanks for the time, man. Yeah.

SPEAKER_02

Thank you for watching and being with us today on this conversation. Don't forget to like, subscribe, and join our channel.