Latest / What's Working in Auto / AI Beyond the BDC with Dustin Schuler, Marketing Director Bud Clary AG
Transcript
- speaker-0: Welcome back to What's Working in Auto. I'm Jake Hughes, and today's guest is Dustin Schuler of the Bud Clary Automotive Group, 14th store group based out of Washington State. Today we're talking all about AI. Dustin just got back from Claude for Dealers event in Atlanta last week. So he's going to go through what he learned at that event, what he's taking back. He was a speaker at the event, so he was working with dealers to educate them and teach them how to go about their process. Dustin is on the cutting edge of AI within automotive. He's using it all across his role, the entire company is a real advocate for the technology and using it in new and unique ways that I think a lot of people can learn from. And that's why I was interested in having him on the show today. Three interesting things we're going to talk about today. One, his approach to monthly advertisements and offers specifically The stores across this group and how he's cut down the production time of that from roughly a day and a half to 30 minutes. To what's possible with AI, how you could build technologies as advanced as a CRM, and why, in some scenarios, he still won't pursue building his own things. And in many scenarios, he is pursuing building his own things for the group. And lastly, how he got into it, how he started, and kind of how the progression has been on his journey to becoming an AI expert and what you can learn from him. So let's get into the show. Welcome to What's Working in Auto. I'm here today with Dustin Schuler. Today we're talking all things Claude. So Dustin, you spoke last week at Claude for Dealers, an event in Atlanta. How'd it go? Give me a little bit of a rundown at the event. I wasn't able to make it. speaker-1: It went great. it's different than any other event that I've been to. So, you know, I've been to many different conferences, NADA, MRC, you know, you name it, I've been there type deal. But especially once you've been in auto for a long time, you know, you don't have yeah, there's a lot of sessions that you'll you'll hear and you'll see and you're like, I kinda have a good grasp on things like that at a lot of these conferences. But this was a workshop where you were building things you could bring back to the dealership the same day. So Mm-hmm. It was a different event, but learned a lot. A lot of people, you know, had things they could take back to the dealership, implement right away, save money or else grow. so I think it was a great event overall. speaker-0: For a little background on you, what's your role? Where do you work? Just give us a your profile so people know where you're coming from. speaker-1: Yeah, I'm the marketing director for the Bud Clary Auto Group. They're out of Washington State. fourteen stores and growing. I think it started ten years ago. We had eight, so we're on a decent pace of growth. and we're still I think we're planning on another three dealers, hopefully within the next year or so. So definitely a growth mindset. you know, for the autograph itself. And then I've been using AAI since I guess raro GPT three whenever that release came out. And then I even helped train some of these AI models with a company called Outlier. It's a subsidy of Scale AI, which is now owned by Meta, I believe. Meta. speaker-0: Mm-hmm. Meta meta. Yeah. So today we're talking about AI. You, in my mind, are sort of the the power user of the moment right now within automotive. So I want to get your take. learn about your journey, your perspective. so the event last week was about cloud. Are you primarily a cloud user or what's your tool of choice these days? speaker-1: I'm primarily a cloud user, at least harness wise. but I use other models. So I don't have, you know, I I don't have like a a set one model that I use. So depending on the use case, you know, images, I might be using Chat GPT. their image generator is pretty good. Video, I'm using Higgs field AI. And then I have a stack of offline models because as of this morning, you know, I'm six days into the month and I've used five point seven billion tokens, which is a lot on Claude, but you know, with your different you know, sub agents, different models you can use within the the AI itself, you're able to kind of budget it where it's not spending a bunch of money, at least. You know, you could use sub models and things like that. speaker-0: So your use has gotten past typical use of what can what can be done with just the Frontier L L so now you're looking for savings. So let's go back, let's say, a year and kind of what's your progression been like in twenty twenty six and what are some of the key use cases that you're finding a lot of value in right now? speaker-1: So I I'm a little bit different. I mean, once you hear marketing, especially for the other people that are listening, you know, the marketers, they understand that marketing means that you kind of do everything. So I would say primarily a year ago, I was doing a lot of marketing specific AI use cases. So think about like automating specials, banners on your website, things like that. to now where, you know, I'm helping with different AI applications for workflows in different departments and also applications to kind of, you know, replace some of the vendor bloat that we have in terms of, you know, I guess the best in a class for one specific use case, but not like an overall system. You know, we haven't done anything like a DMS project or CRM, you know, something that's really big. So we're just kind of doing these ones and twos that are small that, you know, we wish we had this one thing. We wish we had. If it could only do this thing. speaker-0: Yeah, okay. speaker-1: and so especially at our scale, it just makes sense. If it's a five hundred dollar widget, you know, two hundred dollars here and there, you kind of have that death by a thousand paper cuts. So we're like, if it could save t money or save time, that's kind of what we're prioritizing. And then just, you know, working our our way back in terms of like having some easy sprints and then having some that are a little more long term but with a bigger payoff type deal. Just so we have those quick wins, the dopamine hits, and then keep going. speaker-0: Yeah. Do you have a team that y you manage or are you i are you act acting independently? speaker-1: No, I have a I have a team and so you know, it's really important, especially on the marketing side, my marketing team. Even when GPT three, three point five, you know, I guess Claude Opus, probably right around four when that model came out, like those things, I had to really talk to my team and be like, This is this should be a product like an enhancement for us. It should help us move more, metal or else faster, be able to do more things, things like that on the marketing side, because I don't want it to be a replacement for our people. And so especially when I manage, you know, marketing managers, social media specialists, things like that, where, you know, may maybe they have a career after Bud Clary or they want to move on to something bigger. I have to prepare them for that, you know, the modern workflow, the modern, you know, what's gonna be coming up next. so it's a lot of training on that side. But a lot of projects have been self-built. I do have a team, we have an AI committee within our auto group. So, we kind of break up work, you know, where it makes sense, bandwidth wise or expertise wise. And then we're starting to get into the multiplayer AI. So Claude Tag was released in July or August, where you can work basically simultaneously. so it's not like you're working solo, you know, you have your context, you have your basically your claw instance, but then you're able to work with someone else's. So you could both be contributing at the same time. If someone goes on vacation, someone leaves the group, it's not just sit a sitting duck type deal. So you can kind of keep up or keep working through it. speaker-0: Yeah, that's that's been a big problem this year, at least in in my world, same as you. For sure. You you like get some momentum with AI and then you realize the your team can't it's it's not as collaborative as working with, you know, Google Docs or something like that in the past. So can you break down for the audience some basic use cases of productive steps that you've made in recent months with multiplayer and kind of working collaboratively with your team just to make it a little bit more real. speaker-1: Yeah, a hundred percent. I mean a lot of it, you know, it's kind of it's just an accelerator in terms of like if someone say I'm in a meeting or something like that, someone could kind of chime in, help out Claude with the human in the loop situations. But I would say for the mass market, you know, the biggest thing is also getting people that are not necessarily building, but they have wealth knowledge in this specific area that you're trying to, you know, help automate or something like that. And so Mm-hmm. Having something like a tool like Whisper Flow or else Open Whisper, which is a free GitHub kind of variation of it, where you could actually just dump a ton of speaker-0: Sort of tran just like audio transcription tools, right? speaker-1: Correct. Yeah, I should I should have explained that. But WhisperFlow just allows you to hit a hotkey button on your computer, be able to talk into it, and does a really good job of dictation without cutting you off as if you know, you're using Claude on your phone or something like that, where it'll it kind of cuts you off mid sentence, things like that. But allows you to really dump in just a brain dump of context. So this is what we do, this is what we don't do, this is what it should look like. Here's some things we've tried in the past, things like that. Where I mean it just speeds up your learning curve tremendously, especially as you're accelerating, you got multiple projects going on, things like that. speaker-0: Mm-hmm. Yeah, Whisperflow's good. It it's surprisingly accurate and then you just start monologuing at it 'cause you know it's gonna capture everything and you can deliberate with yourself in real time. It it makes it quite easy and quick. speaker-1: And then I can I could also say scratch that and you know, you could kind of redact something like, Okay, I meant to say this or whatever and it just makes it go through clean, which is definitely valuable, especially some of these models, but speaker-0: So are you sitting down with experts on your team and essentially interviewing them to capture the context you need to tackle a specific project? speaker-1: More and more. I think that's kind of the direction that you know, most people should go, especially I don't know everything. I could desk a deal, I could sell finance product, I could I could do a lot, but I don't have fifteen, twenty, twenty-five years of experience behind the desk. I haven't seen everything. I don't know everything and I don't claim to, but so that's it's a really, really good thing. I mean it's it's kinda like the saying of like two minds is greater than one. Mm-hmm. Where you have someone just, you know, brained up and then maybe you get another general manager, you get another service manager, you get a whatever, and you get their context too. And it's like we just made this amazing product, especially at our scale with you know, we basically have a virtual twenty group of different stores within the group. So you can imagine like everyone does things a little bit differently, but speaker-0: What is an example as specific as you can go of a workflow that you've built this year that's automated one automated one specific thing? a tangible example people can kind of wrap their heads around. speaker-1: For sure. so one of the biggest things that I face on the marketing side, especially, you know, I've been doing this for about 10 years now, where you kind of have to do the same go-to-market every 30 days or so. And so one of the biggest time sucks has always been billing banners for the websites. And so we have it fully automated where it's able to pull the incentives, the specials, build them out, place them on the website directly without doing it and across all my stores. It takes about thirty minutes to be able to do that whole workflow. And then I have lead forms on those banners and specials too. So really trying to move to an optimized experience, like a user ex UX type design, similar to Amazon and all these other e commerce shopping businesses, just 'cause that's what customers are familiar with, you know, as we get into the car buying journey and such. and this has been able to accomplish that. But You know, if it takes me thirty minutes, they used to take a full day, a day and a half, especially just doing three banners, let alone doing, you know, fifteen different specials or whatever that looks like and having that fully available and crawlable by the AI and the in the Google crawler. So that's like the tangible example. Probably takes an hour or two to build, but it would save you probably a day, a day and a half, or your graphic designer's, you know, time. speaker-0: Mm-hmm. What's the the flow in your tool stack to pull that off? speaker-1: I mean, right now we just use the Opus five point five and you know, you I have you know, kind of like some database files in terms of compliance and things like that for the different OEMs and then you know, brand colors, brand guidelines, stuff like that. But honestly it's all end to end workflow straight from Plaud within the I'm not even on the terminal C L I. I'm just on the back end of the Claude Desktop application, just running that and given that I I have one password. It's a password manager type deal. And so it's able to kind of pass that information on the cloud so it can log into websites, things like that without me needing to be there. So six AM, as soon as the sensitives come out, it could run that you know, every thirty days. speaker-0: Yeah, I mean that that's fantastic. That must save you so much time and just what has that that enabled you to do now that you don't have to spend that time every single month? speaker-1: I mean it allows me to work on other AI projects and then also Yeah, but then at the same time, like, you know, you have time back in your day, so you know, a lot of the human and loop moments, things like that, where I you know, you're not having AI necessarily do a general manager call to go over your advertising spend, your budget for next month, you know, what you're seeing, what they're seeing type deal. So it frees up time to have, I guess, more collaboration, other, you know, get things done that maybe you know, you would have pushed over, you know, another week or two or something like that 'cause you have a go to market strategy going on. So speaker-0: What were some moments last week at the event that seemed to be significant unlocks for the audience? What did anything specific stand out as like this you explained this thing and now this whole world is open to me? speaker-1: Absolutely. So I was a speaker in the event and before my my session was on Friday. So this conference had been, you know, we had a Wednesday kind of mixer event type deal to you know kind of break the ice, know everybody, meet people. Thursday was our first billing day, and then Friday I had about, I guess, four hours before my session. So between Thursday and Friday, I had a lot of seat time, like actually walking table to table and meeting with the people, helping them get through their issues. And I would say 'cause they split it out in a beginner and advanced track. And so the beginner and intermediate side, a lot of people were net new users to Claude or else AI in general. And so a lot of some of the event like the things that we were doing kind of required you to know how to use it or something like that. So I was able to go through and actually show people easier ways to accomplish things, and kind of have some unlocks in terms of, wow, you know, we could have Claude actually use our computer, our Google Chrome tabs, whatever it is, to be able to do this for me instead of copying this full prompts, whatever. Just tell Cloud to go to that website and execute, you know, step one through five, whatever. And so it just blew people's minds 'cause they're like, man, I could probably have it do this, this, this. I could have it go pull my co op dollars. I could have it, you know, look at this. So, I think that like the biggest unlock. And then on the advanced side, I would say personally, this was it was very interesting to see how many people are building the same things, even though we're like across the country. And so a lot of people were building, you know, a full-on customer data platform, a marketing oper automation platform and stuff. And it's just it was it was wild how how drastic like the learning curve of people that were like I was doing this six months ago and then now I'm building a full platform that we're gonna be using and saving money within our dealership. And so it was it was it was pretty awesome. speaker-0: For those that are building customer data platforms, what are problems are they trying to solve with that? speaker-1: I think that with the customer data platform side, especially some of the smaller groups, you know, they didn't have it in the budget to even, you know, buy an off the shelf product, I would say, or else, you know, maybe that three thousand dollars, you know, could be better used elsewhere. Maybe that's OTT, whatever the you know, marketing spend, or else just helps them become more net profitable. And I think the biggest problems that they were trying to solve, at least at the conference, was how do we protect and safeguard this information? you know, what kind of things are other dealers doing. And that was a lot of discussion points. And so, you know, going through how you become you know, like bring it offline so that it's all intranet, things like that, or using offline models, how do you automate some of the stuff so you're not using your cloud API? I think I think that was the biggest thing, especially on the C V P side. speaker-0: Let's zoom out a little bit here. where do you think most of the industry is at in terms of adoption? Where do you put yourself compared to most people? And what's your best guess at the adoption curve of technology like this? It's gonna be months, years, where are we at? speaker-1: based on what I've seen, I would say right now I would probably be in the top one percent for sure. you know, even with the advanced track users at that conference. I mean, some of them are wicked smart in terms of, you know, backgrounds at, you know, Microsoft, things like that, where, you know, they have the chops to be able to do it. but I would say that learning curve between beginner to advanced to Building maybe application softwares or else advanced workflows, automated loops and workflows itself. That's becoming really short. And so, especially as these models come out with you know, different ways to do it to make it easier. I mean, even we can go with Claude Opus five point five. Once that came out, your prompts had to be shorter. You're not they made similar to Fable when that came out as well. You know, if you put in too many guardrails, you actually get a worse response. ChatGPT came out or open AI, they came out with dots. And so that allows you to, you know, work with a bunch of little sub agents and makes it a little bit user-friendly and easier to adopt. Same thing with Grokbot and I would say Muse as well. and so it's, you know, you quickly become an AI orchestrator for these different agents that you built or that you're building. And I think that adoption curve is becoming very small once you get over the hurdle of just asking it a question or telling it, hey, this is what I do. How can we make it better? How can we make it easier, faster, whatever that looks like. And I mean the models are so good now that you don't have to have a programmer background. So I think the adoption curve within the next six, twelve months, I think we're gonna see I think we're gonna see some wild stuff, some really cool stuff, especially with people that are like, I have n I barely know how to use my computer, but I could use this AI. So think it'll be really fast, faster than people expect. speaker-0: What do you think the industry is trying to achieve by lever leveraging this technology? speaker-1: Well, I think that, you know, a lot of especially in automotive, a lot of things haven't changed. And so, you know, in terms of net profitability, in terms of cars per salesperson that's sold each month, things like that, a lot of the benchmarks haven't changed. but our spend has definitely increased. And so I think that we're seeing, especially with the softer market, you know, affordability pressures there, vehicle cost is higher, we got tariffs, things like that. that was kind of the time, especially with Q4 and stuff like that, where people are like, okay, how do we cut down some of our spend? How do we remove some of this tech bloat? Things like that. Like we have software that people aren't using. We have people, we have duplicate software that does a lot of the same things, stuff like that. and so I think that the industry at a macro is really looking at this as an opportunity to be able to kind of consolidate some technology, have some savings and become more net profitable or to reallocate to grow because you have a lot of small and mid sized groups are trying to grow and you have definitely some mega groups that are, you know, buying up ten store groups, fifteen store groups. So, you know, if anything, if you're able to kind of have this playbook of technology, buy a store, instantly make it more net profitable and sell more cars, I mean that consolidation's gonna accelerate. speaker-0: Do you think there's a play make mid sized groups more financially attractive for acquisition? speaker-1: I think more attractive for acquisition, but also to be able to put themselves in a position to acquire other dealers and deal groups too. you know, I was talking to someone from Russia and or now he lives in Belarus, where he was telling me he plans on coming to the United States, building or buying a dealership, and then instead of you know, waiting three to five years to be able to flip it. it be maybe a shorter cycle with all this AI technology to be a year and a half, increase sales, increase net profitability, and then sell it to a prospective buyer, especially if valuations are high. So I guess what I kind of see on the roadmap. speaker-0: What's been do you attribute the increase in costs in recent years to anything specific, any subset of the of the strategy? speaker-1: I would say that we are I mean we see cost increase you know, every every year or so, you know, there's always like a subtle three hundred dollar, five hundred dollar lift. But then with the vendors that you you partner with, you know, if that happens five years in a row, now your bill looks pretty significant. And same thing with You know, I mean it's the same thing on the dealer on the car buying side too. You know, the cars are getting more expensive every year. Same thing with our technology, our lead providers, stuff like that. So it's definitely a way to be able to kind of cut down on some of that cost. But also a lot of times you run into issues where you're like, I just wish it had this one extra thing. I wish I could do this, and then that's on our roadmap, but that's Q five, you know, a lot of the vendors and stuff like that. So but there's a lot of good things happening in the space. I would say that even some of the legacy companies, a lot of the giants, they're coming out with MCPs. They have full API access at this point. and so we're we're kind of moving into the n you know the the new era where even some of these older companies that I would not have expected to be on the forefront are are making sure that they are with the times. So it's really an exciting time to be an onboat, especially if you're if you're using AI to your your benefit. speaker-0: To what degree do you think all vendors need to be have an AI focus in the way that they build their technology and present themselves? speaker-1: I think it's important, especially you know, if there's an efficiency to be gained or else cost savings to be able to, you know, be had, especially even as a vendor itself. But I think that they also need, you know, vendors and anyone that deals with tier three dealers or even other industries, you have to be able to allow your end user to interact it with it with their AI. So the C Ps, the APIs, the C L I installs like It's gonna be required, especially as people, you know, we call it A D D, another dang dashboard, we'll say. And you know, you have so many different dashboards you're logging in. my wife used to be a service advisor, she's a service manager now, where it's like she had seven different screens she had to log into. She had two different monitors, but you know, you got this tool for texting customers, this one for inspections, this one for your service schedule, this one for your DMS. Like it's just And then you your BG, you get your your training set. Like there's just so much that everyone has to log into, especially like if you're on the corporate side and you have multiple different OEMs, I mean you'll you'll understand the pain where you're like, I gotta do it for this one, this one, you know, you just have so much. So it allows you to more, you know, focus and stuff like that, especially to get consolidated down into one internal tool. And then on our side, especially like when I talk to my IT people. That also allows us to have better control for like SSO. You know, if someone was say someone leaves a group or something like that or onboarding, it's a lot easier to be able to, I guess, turn off access or add them to it, given that it's consolidated down to one dashboard or a couple of different things versus fifteen different deals. speaker-0: Mm-hmm. What advice do you have for vendors in today's market? How it's it seems like cost savings is a big a big focus of yours and and a driver. in the world where you can build a number of things yourself, what are you looking for vendors for support with these days? speaker-1: well well with vendors itself, I would say, honestly we look at it as like risk too. So we do I wanna have the risk of I I could build the CRM, I could build a DMS. That is no problem. I could probably do it within a a day or two. It'd be fully functionable. Do I want all the complexity of that? Do I wanna have the tech debt, you know, if something goes wrong with that, I want I wanna have the SLA, I wanna have, you know, the uptime guarantee and stuff like that. So We're looking at it for security to know that it's done correctly, you know, securely, and it's gonna be there when I clock in on Monday morning or Saturday or Sunday when I'm not there at the dealership, stuff like that. So, you know, we we look at it as like kind of a risk management play. We also look at it as an accelerator because I got I have probably eighty different things on my AI project list right now. Like there's you know I do I wanna tackle that? No. I'd rather just pay for it to be done with it, have the best of that best in class solution for that one specific task or whatever that looks like. you know, and there's definitely a benefit for that, you know, just being able to have not only the speed, but also the guarantee in terms of security and that it's gonna work across, you know, the whole organization. So that that's kind of what I look for right now. And then also as I'm buying or switching vendors, Like for website vendors, for instance, you know, it's important to me that I have an MCP connection to be able to have speed, you know, and be able to interact with it through my whatever AI agent of my choice and stuff like that. So definitely looking at that, especially as I'm considering other products. speaker-0: MCP is now a requirement and everything you're looking into. speaker-1: Yeah, that it we it was a nice to have and then now it's like definitely it it it's becoming more of a requirement where you're like, Man, I I can't imagine having to go through and monitor my staff pages anymore or something like that, even though it's small but it takes me away from something else I can be doing that's profit boosting versus you know, kind of just managing something that's you know, takes me five minutes to be able to log in and remove this person, add this photo, add that, you know. Settle for speed for sure. You want MCPs. speaker-0: Yeah. I'm the same way. It makes it makes a ton of sense. It's just once you get used to it, then it's so much faster, so much easier. Works incredibly well. speaker-1: Especially for being able to pull in stuff for reporting and things like that. I mean I'd rather, you know for a third party, I'd rather pull in the analytics directly versus log into a dashboard and go and click a report for every single store and do that, you know, twice a month or something like you know, it's just for speed. It's so nice to have. speaker-0: Yeah, I find you miss things too. It's hard to have to to not miss some details when you do that. And the AI is very good at making sure to grab every nook and cranny and corner. speaker-1: For sure. But also when you're using the MCPs or the APIs, or else, you know, even CLI, you're able to notice like you have all this data. And even like I said at the beginning of this podcast, where I was like, I don't know everything. I could use the AI to, you know, what am I missing? What am I not thinking about? I could even, you know, kind of brainstorm, you know, look at all of our data points, stuff like that, which man, there's so many different key findings where you're like, man, I wish I knew that five years ago. Yeah. speaker-0: Yeah, totally. I find your perspective to be unique today. I don't think it'll be unique next year, but I think your perspective is unique right now. For a lot of people in the industry I talk to, AI centers around BDC, off hours BDC. If you were to make a pitch to the industry to help everyone expand their thinking beyond that of how valuable AI can be. what are the use cases for AI that you are particularly fond of that are not maybe even directly communicating with customers? speaker-1: There's so many. I would just give you the a framework really to work within. And it's like, doesn't matter what dealership position you are, what role you have, think about what you do every week, every month, and talk to it. And let you like especially with the Whisper Flow or the like, where you could just dump in a bunch of context and say, This takes me eight hours a a week, you know, whether it's you know, something small or big, and you're able to kind of find, okay, AI will actually coach you through what it can do, what it can't do, or else how it could be possible. And maybe that might look like it takes on half of the workflow. And then you have to pick up the rest. But if it's an eight hour session that you have to do to you know, accomplish whatever task you might have at hand, now it's four hours. And then maybe that extra mile of those four hours could be accomplished with the next AI implementation or else the next release from a vendor, things like that, where that's where I see a lot of efficiencies being gained. And you have to think about, you know, what is your time worth? Especially, you know, some of these positions. You know, I get on a phone call with the owners of my company and I'm like, if we had a timer for how and showing like how long this meeting is, and also a money counter for like how much our salaries are at this meeting, we would not have this meeting. Like it would be so expensive. And so, you know, even if It's a a service manager, service director, things like that, being able to pull reporting or whatever that looks like. I would just say that's how you should go about it is talk to it, tell it what you're doing, how it can help, and just let it work for you. So I think at once you start doing that, you're gonna be addicted at the same time. Like you're like, I could do it for so many other things. speaker-0: It's good advice. I think everyone at some point has that moment. Everyone who's a power user now has had that moment of like, Maybe if I just ask and see if it can do this and then it does. And then you're like, Okay, it's off to the races now, throw some gas on it. speaker-1: Yeah, make a recurring task, a scheduled task and you know, don't have to worry about it. You know, send me alerts if something goes wrong, things like that. Ping me on my phone, whatever. Yeah. speaker-0: I've got one more question for you. I'm curious about so you mentioned Muse, Grockbot, all these consumer AI technologies that, you know, for the most part are very new, have come out within the last month. And I think the unlock has been, okay, let's take the concept of agents, proactive tasks on the behalf of the user, make it zero setup, super easy. low cost in a number of scenarios and let let's make this accessible for all. I'm wondering what the consumer AI moment is for auto. Cause I I look at you and I see somebody that is, you know, equ the equivalent of a a programmer in in other industries, just as power user high adoption rate. And I think more of the industry will get in line with you. But then there's so many people working in other parts of the business that Need a more accessible technology. So how do you think this concept of consumer AI could apply in the space of auto? speaker-1: Well, I think about it from multiple different angles, but from the dealership side, interpersonal stuff, I see those consumer agents being really helpful, especially given that you have a enterprise or, you know, for like Claudant, for instance, they call team, where you have like a the safeguards in place to protect your PII. I think that makes it really approachable, especially for people that are afraid of AI or they haven't used AI in the past. to be able to have full agents because you kind of skip from a you know going from a a a chat bot, an LLM, a reasoning model, agents and workflows straight to agents. And so your learning curve is is pretty extraordinary. Like it's it's really short. So I think I could see that, you know, helping for a lot different positions. You know, right now we use like co pilot for a lot of different you know roles that maybe don't need Claude or Chat GPT, things like that. But, you know, as these agents get better and we start seeing it for enterprises, I think we're gonna be scaling it out to where everyone has kind of like the personal assistant, you know, something that's able to understand red flags, be able to bring them up to you, or else do things on your behalf without you having to think about it because you've set up that context in the past. And then from the consumer side, you know, it's definitely a big deal. Like we, you know, I have a site that hit like 830. LLM sessions last month. And you know, so we're seeing already people are buying cars with it. They're at least researching it. Definitely in the research phase, you know, what car makes sense for me type deal. And then they kind of get to, you know, based on if you if you try it out yourself, you'll notice that it'll actually prompt you like, do you want me to send a message to this dealership? Do you want me to find local dealers that have this unit or I found this for you already, just without even prompting it? But in twenty twenty four, we sold our first car directly from AI, where we had the customer came in, they put in, you know, find me the cheapest lease special within a hundred miles of Seattle. And then they it gave them the prompts to be able to send text messages to us, things like that. And the person actually brought in like the full transcript, like emailed a whole a full PDF of it and so showed us. So consumers are definitely using these these models and it's gonna get even better for them in terms of just easy to do, easy to find, you know, it has all the context from them if they've been using in the past. So like I know you have two kids. I know you go on these soccer trips, whatever. I don't think that sedan is a good fit for you. You know, the SUV, you know, go this way to a Toyota Rafael, whatever that looks like. So and based on your driving habits, you know, I know you you work forty miles away. It makes sense for you to have a hybrid. I wouldn't do an E V. Because of your road trips, this and that. Like, so it's it's it's definitely helping on the consumer side. I'd be really interesting I'm really interested in to hear from some of the tier one, you know, the OEMs themselves and how they would be using it. Because they have so much information, especially from our side too, you know, the tier three dealer websites and then also their websites, just to be able to know, you know, at a macro some of these trends that they're seeing. but I I do know that a lot of them are using L LMs and AI models to you know help supply chain issues or revolutionize, optimize, you know, similar stuff. There was an OEM that I I won't name, but you know, they were working on spreadsheets in a certain country and stuff. And so, you know, being able to optimize that, bring it to the twenty first century. I guess the the AI renaissance, you know, definitely helps with some speaker-0: Whatever we want to call this. speaker-1: Yeah, yeah, exactly. speaker-0: Awesome. Dustin, thanks for your time today. I appreciate having you on the show and sharing your AI wisdom with everyone. speaker-1: Absolutely. I hope it helps some people out. If if you you know have any questions, you could reach out to me on LinkedIn. just search me up. But thank you, Jake. Appreciate it.