A Fresh Perspective on Automotive AI Strategies

Brian Pasch0:00Hi, this is Brian Pasch and welcome to another exciting podcast episode. DMSC is over and the buzzword, whether justified or not, is AI this, AI that. So I decided to bring automotive industry veteran, AI leader, pioneer, and one of the few people who actually has thousands and thousands and thousands of examples of AI in automotive, Albert, CEO of IDPrivacy.ai. Hi, Albert. I'm so excited to just have a conversation to help dealers get up to speed. So for the dealers who don't know you, give me a little thumbnail of your work, your tenure here in automotive.

Albert Thompson0:57Yeah, absolutely. And Brian, thank you so much for the opportunity to be here. You know, I really, I've always valued everything you're doing and talking about a pioneer, I feel very honored to be in front of a pioneer myself. So thank you again.

Brian Pasch1:08Thank you. You're very welcome.

Albert Thompson1:11So Albert Thompson, as you mentioned, CEO and founder of IDPrivacy. My background has been, I say 25 years, Brian, but it's been longer than 25 years. I started when I was 16 in automotive and haven't been able to escape since then. Sold my first franchise store in Northeast Ohio at 16 and Carfax in the early 2000s and Autotrader. And from there, you know, the background just continues to keep layering on, you know, went to help launch TradeRev and took that to KAR Global when that was exited and was one of the first few to also help launch Drive Auto. Took that through the Sinclair Broadcast Group and built their digital tech stack till I actually launched my previous company, which is, Brian, we spent a lot of time together on that side, which was the first party identity layer.

Albert Thompson1:59And I was able to take an exit early 2024. And now here we are with IDPrivacy.ai.

Brian Pasch2:06Great. So Albert, I've been following you on LinkedIn. First of all, you have the most thoughtful posts. They're not short in any way. You love to write like I like to write. And sometimes I wonder, should I make this short or long, but I can't help myself. I wanted to interview you because there seems to be a mad rush for AI solutions. My concern, and I've brought this up a number of times, is who's controlling the consent, who's controlling the cadence, and who's controlling the context if there's three, four, or five different AI platforms being used in the dealership. So give me your snapshot. What are you seeing?

Brian Pasch2:56Obviously, you're knee deep in all of this or neck deep. But what's your observation in how the vendor community is approaching AI and selling those solutions to dealers?

Albert Thompson3:11Yeah, you know, it's interesting that you actually brought this up. It's one of the biggest pain points and one of the biggest problems that we, early on, when I launched the organization, saw as a problem, right? And if you think about it, literally the thesis, the problem statement, the thesis that we saw in the very beginning, especially as to how it pertains to AI and how it's going to, you know, integrate within a dealer operation, is we already have system sprawl and we already have data sprawl, right? We have average dealer, I think, has nine to 12 systems today. And here's the problem, right? Now we're talking about adding AI tools on top of tools and on top of systems.

Albert Thompson3:50None of them are talking to each other. None of them are interoperable. And I think you said, right, consent. None of them have, they all have individualized consents. So there's no unification to this. So my biggest problem, my biggest fear is that we're just going to further fragment AI on top of further fragmented systems and ultimately is going to lead to a messy divorce with dealer and AI tools in six months.

Brian Pasch4:15Yeah, and, you know, we're going to get into that and what your company is doing, but what do you think one of the biggest misconceptions about AI is for dealers? Because it seems to me that dealers aren't concerned right now about consent, cadence, and context. They just see it as, yeah, that would be good, but this guy's helping me with my service. This agent is helping me with lead follow-up. This agent is responding to phone calls. What's your take on the biggest misconception with AI in automotive?

Albert Thompson4:57So I think there's several misconceptions, and I think you just nailed it, right? The fact that we're thinking about agents in the sense that I have an agent for service, and maybe I have an agent that's going to do my SMS, and then I'll have an agent that's going to do my SEO. I think that's the first misconception. The biggest misconception is that, you know, we think the AI or agents is something that you just buy and that you just bolt on. And I get why that misconception exists. I totally get it. Dealers have been buying software their whole career, right? They know how to evaluate a DMS.

Albert Thompson5:29They know how to negotiate CRM contracts, and they know how to kick the tires and watch the demo and sign the order form and implement. That model works for static software. Unfortunately, Brian, AI is not static, right? It's not deterministic. It changes. It learns. It adapts. And you mentioned that with the cadence. And to me, cadence is a big word for orchestration. And if you don't have a native AI stack that works across service, sales, parts, finance, and the entire dealer ecosystem, that cadence doesn't work, right? So now, like I said, you're going back to further fragmenting it. So misconception number one is that AI is not deterministic.

Albert Thompson6:10It's probabilistic. And so I don't think the dealers quite understand that. You know, they say, you really have to understand that you don't buy it like software. It doesn't implement like software. And it can't just be bolted on. That's misconception number one, right? You can't just take one tool here and say, that's great. That's my service voice agent. And I'll do this one for my sales. And I'll do this one for my SEO. That's a problem in itself. Number two, the other misconception is that dealers tend to think that AI, for some reason, we've all thought that AI should be perfect. And I don't know where this came from, but I think everybody thinks that the agent should be perfect from day one.

Albert Thompson6:49And I explain to dealers all the time. I have many conversations. I say, listen, if AI agent, as you say, is perfect from day one, it's not AI. It's scripted. It's automation. And scripts break the moment the customers go off that context you're talking about, off that cadence you're talking about, and say something that the vendor or the scripting tool did not anticipate. And that's that orchestration layer. So in automotive, that's going to be every three calls. You know, my dad actually owns the car. I'm calling because my truck is leaking, but I also want to trade it in. Can you help me do that?

Albert Thompson7:20Real agents can handle those scripted ones, right?

Brian Pasch7:23Right.

Albert Thompson7:24And the other thing is, too, right, is we think that these AI agents are going to be perfect from day one, but they have to actually learn. And I think there's a third misconception, right? And this is an important one for me. Mistakes in software means failure. And that's true, right? So if you buy a website and it's not loading, you have an error 400 or 404, that's broken, right? That's true. Mistakes in AI does not mean that the AI failed. Mistakes actually mean the system's alive. If your agent never makes a mistake, it's not AI and it's not learning. And this is important because dealers have to go in with a different psychological mindset and they have to go in with a different understanding of how this is going to deploy in their business.

Albert Thompson8:07And what you should measure is not whether it makes the same mistakes or makes mistakes, but can it learn? Can it get better? Right. And do these become teachable moments? The mental shift I ask dealers every day to make, you know, you're not buying software. You are hiring a little team member. You have to train it. You have to coach it. You have to nurture it. Give it some guardrails.

Brian Pasch8:30Yes. Coaching. Yeah. So think about that for a moment. Dealers have primarily the same business model, but how they execute is very different. So when you think of where the industry is heading, do you think that dealership executives now really have to get some training education on how to train these agentic agents with their mindset, with their protocols, with their processes? Now, I've heard vendors say, well, just give me your employee handbook, give me your training manuals, give me, you know, whatever, and it will just absorb it and learn from that. That seems very optimistic for me. So what expectation would you like to set, especially on this third misconception?

Brian Pasch9:29AI agents are being deployed. How much work does the dealer have to do in those first few months to make sure that their newbie 24-7 employee is doing the right thing?

Albert Thompson9:44Yeah. Let me start off by taking a step back and painting a picture for us. Let me frame this up. For years and years and years, right? Think about your general manager, right? That phone ringing meant cha-ching, right? That meant dollar signs. Right. Every time they heard that phone ringing in the dealership. So now I literally say this statement to dealers. And now I want you to imagine on a Saturday morning, your sales guys are kicking back. Maybe they're playing on YouTube. There's not a soul on the lot. That phone's not ringing. And your owner, your dealer principal picks up the phone and goes, why is there only eight appointments on the books?

Albert Thompson10:19We normally have 20. And you have AI deployed and the phone's not ringing. What do you do? And in real-world scenario, right, you have this control, right? You could say, get up out of your desk. You go to the BDC. Guys, go pick up all of the sales from the last six months and pick them up and call them. And you know what?

Brian Pasch10:38Right, right. Yeah.

Albert Thompson10:40Give specific tactical instructions because the human observation is saying we have a problem based on historical norms. And we have a psychological need to try to fix it, right?

Brian Pasch10:52Yes.

Albert Thompson10:54What are you going to do? You're going to yell at the AI agent? Probably not, right? So what you'll do is pick up the phone and fire your vendor. Get it out of here. It's not working. Kick it out.

Brian Pasch11:02Right.

Albert Thompson11:03That's the first thing you do because that's the first control you can take. I'll just put my BDC back in or I'll just put my sales guys on this because this isn't working. I had 12 appointments and I'm like, I got to tell you, this is very important. This is not something you just roll in and put in there. This is like deploying a DMS. This is like deploying a CRM. You have to, you know, you have to have buy-in from your teams. You have to have buy-ins from the BDC. You have to have buy-ins from leadership all the way down because these scenarios really happen.

Albert Thompson11:27And there is another side. There is a there there. I've seen it. I've got the case. I've seen it literally happen in every store that just pushes through these moments. But those moments are real. And I think that a real partner to a dealer is going to set that expectation up from the ground floor up.

Brian Pasch11:43I love that. You know, I asked people on LinkedIn a question. The funny thing about LinkedIn, and you probably know this, Albert, as well. When someone's impacted by something you write or I write, due to the politics, they can't always comment. So they'll send me a message like, hey, read your article. Well, that's awesome. I can't really publicly state that. Or, hey, I can't believe people are voting like that. Whatever. So whatever people see on our posts or newsletters is just a fraction of what's going on. But I put up a poll recently asking people in automotive who should be the controller of compliance for AI outreach.

Brian Pasch12:31So if the CRM, for example, let's just say you have a DealerSocket, I mean, excuse me, DriveCentric with agentic agents, and then you have Dealer Funnel with some AI agents doing marketing, and then you have another AI tool that is running, you know, chat on the website, and then another AI tool that's handling service and sales phone calls, right? Let's just say that's very common to potentially have four different people using some type of AI intermediated communication. So I asked, who should control it? Is it the CRM? Is it the DMS? Is it the CDP or an independent platform? And it was funny when some people said the DMS, which is kind of funny because the DMS is

Brian Pasch13:27like a dirty cesspool of data. So I had private comments, people saying whoever said DMS is dumb as rock. Okay. So, but here's the thing that this is just how people vote. It's a reflection of how they're thinking. They, anybody who says the DMS should be the central repository for, you know, unified customer communications hasn't really understood how dirty their data is. Albert, what would you say in a world where most CRMs are going to have an AI layer built in and potentially, potentially those CRM companies are not going to be handling in and outbound phone call. So you might have a second company, CallRevu, Car Wars, Toma, Numa, Mia, Spyne.

Brian Pasch14:21I mean, the list is growing. If you had a future cast, who's going to control the permissioning, right? Consent, context, and cadence. In a multi-vendor world, where should that sit?

Albert Thompson14:43Yeah, great question. And I think, you know, personally, and this might sound self-serving, I don't think it's any of those tools today. I think it's actually a system of action, right? That's purpose built for agents. So think about that for a second. And the CDP, right, is built for, you know, giving that unified, it's built for the marketer, giving a unified view of the customer so they can segment, target, and activate across the digital pathways, right?

Brian Pasch15:07Correct.

Albert Thompson15:08The CRM, system of record, built for the dealer, built for the salesperson, managing the leads. And I can see why a CRM would want to have tools that could help to manage leads. But a system of record ultimately isn't that system of action, right? And so this kind of leads into us, right? This is a little self-serving. But ultimately, I think the engagement operating system becomes the layer between the customer contact, the conversation, and the conversion, right? And in that middle layer, that's where these agents can have context graphs that understand real-time context that's happening with the customer. And all of those, it's that unified layer that would operate across all of the different

Albert Thompson15:51dealer profit centers. And that layer would be purposely built to do what it's supposed to do, which is handle those customer interactions. And I think in a perfect world, I think you'll see the CDP underneath, right? Because you need to make sure that you have that as the foundation. You'll have the engagement operating system sitting above that, right? Where the agents are literally just focused on engagement and customer interactions. And then ultimately, they'll be able to interact directly and out of the systems of record.

Brian Pasch16:21So let's dive into that. For dealers who don't know what your company's North Star is, what your goal is, what's your elevator pitch? Because it has many levels of intrigue. But what would you do to kind of summarize for people on our podcast where you believe your unique skill set and vision is to help solve dealer problems?

Albert Thompson16:51Yeah, so essentially, that's exactly right. So IDPrivacy is an AI engagement, agentic AI engagement operating system. But let me just make simple sense of that, right? Is that we have built that middle layer that brings in, I don't care if it's a third-party lead from Autotrader. I don't care if it's an inbound service call, an inbound sales call. It's the purpose-built operating system that sits underneath of all the dealer profit centers to capture every interaction and activate and manage those leads and those customer conversations end-to-end through sales service and retention workflows using voice, SMS, and email, and chat. And with that, these agents, they are context-aware, channel-aware, and journey-aware.

Albert Thompson17:37So I'll explain that. So, and you see today, and you mentioned some of those tools, right? There's tools that may be just taking inbound service calls. But that falls short, right? Because there's real context in those service calls that are relevant to a sales agent.

Brian Pasch17:51That's correct.

Albert Thompson17:52And those agents need to understand each other. And that context gets lost if you don't have that operating system in between.

Brian Pasch17:58Yeah, and let's just pause for a minute, because I always like to give a dealer example. What Albert's talking about, if somebody calls to book a service appointment, an AI agent may do a perfect job. But they may also say, hey, by the way, I'm thinking about trading in my car.

Albert Thompson18:14Yes.

Brian Pasch18:14And I'm not sure how much these repairs are, so I may consider a trade-in. Well, an AI agent just built for service will book the appointment. When you have agents working together, then there would be a handoff and say, hey, look, this customer's coming in for service, but they're also thinking trade. Maybe start reaching out to them and finding out what vehicle they want to trade into, you know? And what you're saying is having all of those agents working together in an engagement layer is what makes the magic work.

Albert Thompson18:54Yes, that's exactly right. And I think that that's the big difference between, say, like a chatbot and say, you know, great, that's good information. And then it basically just takes it, and it's a passive system. So take that same scenario, right? Somebody calls in for service, you know, and you've seen it today, right? Are you new, returning? And then, you know, hey, I see you have this vehicle on record. Great. And the customer might even say, hey, actually, I sold that car, and that's a big problem for us, right? That car's no longer. So the agent would still just book the appointment and go through it.

Albert Thompson19:25And then maybe the customer does say, but I am still thinking about trading it in or maybe looking at an opportunity. That appointment might just get booked, and it's just passive. That's it. That tool stays static. It's done. It sits onto the side, and that is ultimately, there's just no interoperability. In an engagement operating system, essentially, that would come down, right, to that context layer. And that context-rich response would then pass over to a sales agent who could then, you know, put all that together, actually take out the vehicle that the customer said, I don't have anymore, right? Which then becomes this living, healing, intelligent graph, and then update it to validate the vehicle that's their car, and then update them as a trade intender.

Albert Thompson20:07And then that next outreach might say, hey, Mr. Smith, you know, I saw that you reached into service and said you had a Rogue you were interested in possibly trading in. You know, it'd be really great to talk to you about the new Rogues. You know, do you have 30 minutes this week to get you in? And that's a context-rich approach to the agent reaching out.

Brian Pasch20:21So, Albert, you know, there's a lot of claims about the benefits of AI, whether it's taking inbound phone calls, more recently people doing outbound phone calls, obviously the lead handling 24-7. Again, I always worry about, you know, this idea of compliance, context, and cadence. But you have published and shared online a number of examples in scale, meaning nothing insignificant. For dealers who are really looking for a partner, for an agentic solution, and they want to understand what benefits they could have from dealing with your company, can you give me one example for variable ops, for the sales people leaning in, and one for the fixed ops?

Brian Pasch21:22Because we are getting more and more fixed ops directors listening into our podcast. So, could you give me two examples, and would you also give us the context of how long it takes to kind of ramp up to speed, right? So, you just said, you know, you said earlier, Brian, I want people to know that there's some training. So, when you say, hey, this Toyota dealer book, you know, what was the context? It was that, okay, in their first 90 days, they got everything perfect, and then that fourth month, they rocked it like this. Give me an example for variable ops and fixed ops.

Albert Thompson22:01Yeah. So, you know, so our system, you know, really, you know, first commercial agents that went out are the sales and service agents, right? So, very simply put, you know, our systems are bi-direction integrated into most of the CRMs and most of the service schedulers in the platform, right? So, you would deploy, let's just start with variable ops, right? So, you would deploy an inbound and an outbound sales agent. And I say inbound and outbound sales agent because they're two different things. You know, I've heard a lot of other companies saying, oh, we should have one agent that does it all. And that's actually fundamentally just doesn't actually make sense, right?

Albert Thompson22:32Because there should be agent-to-agent, you know, handling. But as an example, if there was a GM or a dealer principal listening right now and you wanted to actually deploy an agent with IDPrivacy, we would essentially integrate directly into the CRM. We would deploy an inbound agent for voice, SMS, email, and chat, all of them interconnected and talking to each other. So, if somebody comes to the chat and here's an example and says, hey, you know, I'm looking at this vehicle, you know, is this OEM still offering 0% financing, engages, disengages. Maybe the next day they pick up the phone and call, then the agent would actually be able to take that call and know who that person is, right?

Albert Thompson23:14Be able to have that conversation. Hey, are you still calling back about that vehicle with that 0%? And then essentially on an outbound perspective, right, if that appointment doesn't book, because the goal of the inbound agent would be to essentially get through vehicle of interest, identify trade opportunity, finance interest, and then ultimately book the appointment. And also, that doesn't matter if it's from Autotrader, as soon as it hits the CRM, or if it's an inbound call, the agent can work that inbound on an outbound, it's going to work that first quality response. It's going to maintain brand voice, tone, the important things that matter for the continuity of that brand presence.

Albert Thompson23:51And then, of course, just work out outbound SMS, email, and, of course, outbound voice as well. And I'll explain how we handle that from a compliance perspective as well. But in terms of timing, let's just talk about from sales time, and then I'll switch over to fixed ops. You know, what I always recommend dealers is 90 days. We've seen dealers go from first start to, like, record months in 90 days. And I call this the first time to first value. It's one of the KPIs that matters. I mean, this predicts whether or not your deployment will survive or it'll fail. And I think that most of the dealers want to hear that it's going to start right away.

Albert Thompson24:27It does. By the way, don't get me wrong. Day one, within hours, you're going to be booking appointments. There's no question. The outbound agent's going to be making calls, booking appointments. But it's going to have, you know, where it might say GX460 instead of GX460. And this is where we have to understand that the agent's learning. And I call these teachable moments. And we shouldn't focus on the minors. We should focus on the majors. Did we get month one lift? You know, did you do 80 appointments last month? And now we did 96, you know? This is the stuff that we should be thinking about.

Albert Thompson25:02And then month two, you know, did we do better than that? Did we do better than that? Are we closing? And then using the humans, and I want to be very, very open about this. You know, we don't believe that AI, you know, should take human jobs. Humans should, if you were a great salesperson, you're going to be a stellar salesperson if you do this right.

Brian Pasch25:19Right, right.

Albert Thompson25:20You should augment, right? Let the AI do the redundant, monotonous work. So imagine all of a sudden now the agent is working all the inbound leads and the outbound leads, doing all the texting, SMS at scale. We build our own orchestration layer. So let me explain this a little bit. This is very important. We don't use third-party tools like, you know, n8ns of the world or Makes of the world to stitch workflows. That's actually one of our biggest secret sauces. Took us about two years to develop and to get right. But the key is the agents can decide and reason and act autonomously based on context, like

Albert Thompson25:56you said, compliance, guardrails, policies, and then ultimately, you know, the conversations and the signals, and then it can decide what the next step is. So if John Smith says, hey, I'm interested. Can you give me a call tomorrow back at 12? That agent will look at, you know, are they open? Is it Sunday? Are they closed? All these different things, right? The guardrails. And then decide, you know, if that's the best move and then it'll make that decision. So you've got agents working your leads 24-7. Obviously, not calling people or anything like that at midnight because that's the guardrails, right? And then when they become an active hand raiser, then it pushes it back to the human as an alert.

Albert Thompson26:31And then that's where these humans should be focused. You know, focus on when the agent gets them to the point where they're ready to act. Then you engage and bring them in for the showroom. And then dealers that actually follow that methodology are seeing incredible results. And on service side, if they were...

Brian Pasch26:49Albert, before you go to service, I was just thinking something. And I don't know if this analogy holds water. I'll let you react. But a few times, and I'm not a student of the best super salesman in, you know, automotive retail. But there have been some amazing people, you know, who are selling like, hey, I'm selling 80, 90 cars a month. And you're like, yeah. And then, of course, I have to ask, well, how are you doing that? It's kind of like, I have an assistant. And some of these people have two assistants. And what they've figured out is the tasks that they're really good at, that add value and keep people engaged.

Brian Pasch27:36And the tasks that they're really not good at. And by having these assistants, their throughput, instead of 10 or 15 cars a month, or 80, 90 cars a month. Maybe we should be using that analogy more that the salespeople who have a vision to do more, to serve more, to grow more, can look at AI as their assistance to take care of the tasks that are not adding value and focus on the tasks that really bring value. How does that sound to you?

Albert Thompson28:13That sounds like spot on. That's exactly how this should be looked at. AI for the assist. You know, the thing is, and I've even coached dealers, right? You know, I think this is something I actually even mentioned right now. You know, you'll sometimes get dealers and they'll knee jerk, right? Because maybe the AI is answering the calls. And all of a sudden, somebody calls in and complains, you know, this, you're using AI, whatever, right? And you're going to have complaints. And dealers, you know, they'll get nervous. They'll be like, oh, this is, you know, it's an AI. The reality is this, is of course it's going to get complaints.

Albert Thompson28:44If it's doing 100% of your calls in service or in sales, for example, and a human isn't, you're expected to get mistakes. But going back to what you just said, the response, the proper response to that customer will be like, but you know what, Mr. Customer, Mrs. Customer, we value you so much in showroom. The experiences that we want to provide to you when you walk into our showroom and when you're buying a car here, we feel that, you know what, it's better to have an assistant, a digital assistant answering our calls so that we can put all of our time and energy to you when you're here at the dealership.

Brian Pasch29:16Come on, let's go. I love that. Albert, let's talk about fixed ops. For the fixed ops directors or for the general managers who are looking to just get a little better education on what's possible on the fixed ops side, what are you seeing for your customers?

Albert Thompson29:33Yeah, I think there's a lot of things that are happening. First off, let's talk about consumer adoption with AI. I think that's important. These last two years have been really interesting. We've seen millions of interactions. I think we're just shy of a million plus calls, over 130,000, you know, service booked appointments. So we've seen a lot. And I've seen it go from customers immediately, like really most of the time wanting to transfer to a human and not wanting to talk to an agent. To now, all of a sudden, you know, 70, 80, 90% of the calls are booked with the AI agent and the customers are happily to do that.

Albert Thompson30:13And I think what's ultimately happened is the consumer adoption is to the point where now they're like, hey, my time's valuable. If this agent actually isn't going to be wonky or clunky and actually get me through the process and get me through quickly and actually solve my needs, the consumers are okay with it. That is absolutely 100%, I think, to this point there. I mean, you might get one out of 30 that might say, you know, I don't want to talk to an agent now. It's incredible, especially if the agents are doing a good job.

Brian Pasch30:41Albert, I want to piggyback on that. I have so many of my friends, men and women, that are just like, hey, man, I'm using ChatGPT for everything. I was just talking to a married couple. I was like, ChatGPT is the greatest thing. Are you using it? You know, so number one, I'm seeing that. Number two, because I like analogies to bring people back to where we are. I'm 64. So when I first started traveling by air, you went to a travel agent and you got paper tickets and you got them mailed to you. And if you lost those paper tickets, you were out of luck.

Brian Pasch31:19When the internet first opened up, the booking portals for, I was mostly flying United at the time, were kind of clunky and hard. And you would try and you got frustrated. And then you ended up calling. Today, I am the expert travel agent. And I can do multi-leg flights. I'm looking for discounts. I'm looking for seats and plane configurations. I don't want to talk to anybody at United or Delta unless I have a major problem. And I wonder if we're in that transition, right, that over the last year or two, we're throwing up some AI models, hoping to solve some gaps. More and more people are working through them.

Brian Pasch32:01But I still think we're in the early days. And so, like you said, dealers shouldn't give up because the revenue lost is invisible.

Albert Thompson32:14Yes.

Brian Pasch32:14Because you don't know if people are calling. You don't know if people are doing the work. With the AI workflows, you know the work is being done. So, in effect, the calls are being made. The emails are being sent. The texts are being sent. And now, for the first time, you can see, you know, where maybe your process isn't scaled properly. But this is a time of transition. So, what are we seeing in service? What's some exciting news about filling gaps that have been in fixed ops for a long time?

Albert Thompson32:52Yeah. Well, first and foremost, I love what you just said about seeing and identifying processes. And I think this spans across variable ops and fixed ops because for the first time ever, right, you have a different set of lenses that you can operate off of as a dealer. Never in the history of automotive, right, have we actually had the ability to look, listen, and learn from everything that's happening in our dealership at scale. Because not only do these agents take the calls, right, but you're pulling in that intelligence from the transcripts, what the customers are saying, why they're transferring, what the transfer was about, are they upside down on trade?

Albert Thompson33:31You know, you're picking up the signals of, I had a soccer mom the other day picking up her kids, and so you're learning she's a soccer mom. All of that nuance, and so really the technology, the moat, Brian, is not the cause, the technology of the conversations like the text and emails and the voice. The real moat is that data intelligence behind it. The exciting things is that. The exciting things is what you can do with that data. The second part of it is having agents that are working together as a team in the background. Imagine a call comes in, a person says, hey, I've got a Tesla, I want to schedule for a change in my battery, and so I want to come in today.

Albert Thompson34:06Well, that's great. I can certainly make that appointment happen, but maybe that battery's not in stock, right? So in milliseconds, you need to have another agent that's looking in the parts, looking for availability, saying, oh, you know what? I don't have that part, but just sent off an order right now, put the order in for the battery to come in. All this is happening on a call where the customer in seconds doesn't realize it. Agent comes back, says, you know, Brian, love to get you in today. However, your Tesla battery needed to be ordered. It'll be here Friday, April 23rd. Why don't we look at the appointments that day at 830 and 1130, which works for you?

Brian Pasch34:40Yeah, that's sick. Now you're talking about customer CSI at a whole other level. You're not bringing them in there for parts that aren't available, you know, setting proper expectations, and these agents are all working together.

Albert Thompson34:52And then on top of this, this is where I keep going back to that context layer. That's real-time signals. So you know what I love about the CDP, right? The CDP is the foundation. And you and I, we'll talk about this later, but that CDP being the foundation and having that clean data layer. On the operating system layer, right, that system of action layer, you're picking up real-time signals. So now all of a sudden, hey, this guy's got a Tesla. I know he's got a battery. I know he needs to order the battery, right? You know, he's at 43,000 miles. There's equity opportunities. All of that is real-time signals that need to go back to that CDP.

Albert Thompson35:22So it needs to make its way to a context graph. It needs to make its way back to that CDP.

Brian Pasch35:27Yeah, and this is why I wanted to have you on because you have had a jump start on the CDP world, marketing activation. You've been really ahead of the things that people know me to write about. And I saw you were doing all sorts of things with AI and started publishing, you know, dealer success stories. I'm like, no one else is publishing ever. I was, like, already, you know, through, you know, the trial by fire, you know, getting all these workflows to work together. And that's exciting. And that's why I think, like, in the future at MRC, I think it would be great for us to figure out,

Brian Pasch36:14how do we work on something that really educates the dealers, like, you know, something meaningful, like we did the CRM survey of dealers. Dealers really liked that. I'm wondering if we really should be doing some, almost like an AI playbook, you know, what should the dealers be thinking about, like, when I published the guide on CDPs? You know, I think you have a wealth of information. One of the other pieces before we close today, you mentioned it, time to value. When I was working with Tealium on the CDP project for Morgan, value engineering was brought up. And, you know, when dealers normally buy software, that's never brought up.

Brian Pasch37:07And Tealium was very good about let's set up a plan so we can show you the value engineering, what's important to you, what are the goals, and how long is it going to take? And I was like, that's a really good exercise, but you, you also brought it up. So you're in an enterprise mindset. I think I want to dive into that a little bit. When dealers ask you, whatever your platform costs, and I don't think cost is what I'm interested in. I'm talking about how do you explain to them the value engineering piece, like, hey, if you trust me with your engagement layer, with this communication layer, with the agents that we've built, you mentioned in and outbound for sales and service and parts and warranty and recall and F&I, all the agents that you've built.

Brian Pasch38:11How do you talk to them about the value delivered for their investment, right? They have to, you said, management has to buy in. They have to buy in. People have to push through learning curves. We understand that this is an ecosystem that is learning and refining and improving over time. What's that conversation normally look like for a dealer or a dealer group?

Albert Thompson38:37Yeah, great question, right? Because I think a lot of it is discovery first, right? Helping them to understand exactly what this means. Because like I said, I really, and Brian, I really want to echo what you said at the very beginning of the call about these stacking tools on top of tools. We're seeing that today. And I'm nervous to see what that's going to look like down the road. Because I think the unfortunate thing is, yeah, you can get some instant value, right? You can start to see booked appointments. But there's a bigger conversation that you need to be having with the dealer groups and the dealers.

Albert Thompson39:09And that's that long-term, like this isn't going away. It's not going away. It's here, right? So we really need to sit down with them. And I like what you said about the operational mindset, because I think that's how we need to be thinking about that from a dealership perspective. We need to be talking about operational maturity. Anybody can sell you a demo, but we really need to be thinking about, you know, predictable timelines. You know, what does this look like implemented with the BDC team? How does this look like with your sales? So we need to sit down and, you know, what's important to them, right?

Albert Thompson39:40You know, when we talk about first time to first value, they're counting the clock, right? And they need to see measurable value. So what is that measurable value, right? Is that, hey, I need to see my appointments booked, you know, going from here to here? Okay, that's fine. Do I need to see, you know, my data layer, you know, being, you know, more robust and having more signals so that I can activate better? I think all of these AI deployments need to have some sort of KPI matrix. And every dealer is different. I think, you know, we start from scratch with every dealer. And that's the other thing, too.

Albert Thompson40:12This isn't set it and forget it. I found that, right? You can't just set this into any store and say, okay, here you go. Here's your service agent. It's going to answer all your calls and you're set. That's not going to work long term, right? You have to really sit down and do that. So I think the answer is it's every dealer is different, but it's all built off of the same framework, right? We know that you're ultimately have some pain points. And what are those pain points? Because think about this for a second. You have the in-service, for example, you now have coverage. You have bandwidth.

Albert Thompson40:46Before you were bandwidth constrained. You know, Brian, think about this. How many times have we talked to dealers and they just want to buy more leads, more marketing, more leads? Because everything comes to the bottom funnel and that's what their salespeople can handle. They can get more leads and more leads ultimately they think means more sales. We fundamentally think differently. You don't need more leads. You need a better process.

Brian Pasch41:04Right.

Albert Thompson41:04To manage those leads, right? And so that's, you know, resetting the mindset of the dealership and saying, hey, you now have a tool that can do all of these things. So where is that value? It's not maybe more leads. It's how do we want these agents to work those leads? How do we want them?

Brian Pasch41:19You know, Albert, you bring up this more leads, better process. I want to tell a story because I want dealers to remember certain things about our conversation. Until Urban Science started showing the defection data, dealers didn't know if their processes were broken. And I'm going to come back to that. Just as you said before AI, dealers never knew that their process is being fully executed 24-7 and they could see if that process is scale tested or volume tested. But let me go back to Urban Science. When dealers see their defection data and what day the person defected and the context for the dealers here, if they don't have this data, it's the most valuable data.

Brian Pasch42:05If a lead is in your CRM, Urban Science, as soon as that person buys from another dealership, they're going to send you a signal into your CRM. If you have both SalesAlert and TrafficView, they're going to tell you if that person bought from the same brand and was it a local. So imagine you're a Honda dealer, you think you have your processes locked down, and then all of a sudden you see on day 14 through 21 huge defection rates. That isn't a lead problem. That's a process problem. And then my brother Glenn goes into their CRM and says, yeah, your workflows are really, for the first 14 days, you stop in your mindset.

Brian Pasch42:44You know, your process kind of goes on autopilot after day 14 and look at all the people who bought from the Honda dealer down the road. This is going to be a beautiful time, I think, Albert, when you are able to guide dealers using data sources to refine the processes, right? The AI models and workflows we have today are going to change over time. But here's the cool thing. We're going to start to get in these external signals so that we can get better at lowering defection rates, having contextual messages, as you mentioned, cross department. This is an amazing, amazing time. For dealers who are leaning in and saying, this is the first time I heard somebody talk common sense about AI and let's work through the business needs, let's do a value engineering, you know, timeline.

Brian Pasch43:41What's the best way for them to get in touch with you or visit your company website? What's that URL?

Albert Thompson43:49Yeah, thank you, Brian. I appreciate it. It's www.idprivacy.com. Just like your ID. It's actually, and by the way, I'll take an opportunity to say this, intelligence, data, private, focused AI, right? That's it. That's not ID, like a driver's license, intelligence-driven, privacy, AI. So it's idprivacy.ai, and that's where they can find us.

Brian Pasch44:12And what if they wanted to send you a direct note, Albert? How do they send you an email?

Albert Thompson44:16Yeah, albert, A-L-B-E-R-T, at idprivacy.ai.

Brian Pasch44:21That's so easy. Albert, there's so much. We're going to have to put our heads together because now that DMSC is over, we are going to set our sights for November for Modern Retailing. So why don't you put your head together on what you think dealers need most to help them navigate through the AI transformation of their business and to avoid some of the potholes, and then let's work to do a presentation in November. How does that sound like to you?

Albert Thompson44:50No, I think it sounds fantastic, Brian, and I want to leave you with this because I know we're coming on time, but the last piece here that we just touched on, this defection piece, is so important. Because if you actually think about it, and this is a learning lesson, and it's a very exciting time, AI did not break your process. AI revealed your process was already broken.

Brian Pasch45:13Yeah, come on. And this is the exciting thing. Here's what I'm wondering. I'm wondering when we really get these workflows right, the feedback signals from companies like Urban Science and what we lost, what's the limit to the conversion rate of lead to sale? Right now, you know, dealers will say, man, we're crushing it, we're doing 17%, you know, lead to sale, you know, 35% appointment to sale, whatever their, you know, their stats are. It's amazing what the future could be like. I wonder, I only can guess what it could be.

Albert Thompson45:55Well, we just did an OEM case study. I won't say the OEM, I'll see if I can get permission to get you that. But we did an OEM case study, tier one. We managed all the tier one leads. The agent gave 30 minutes to the dealer for the first 30 minutes to respond to the lead. From then, our agents took over, right? And what we found in this particular OEM, now everyone's different. In this particular one, it's a, you know, compact, you know, economical OEM. But leads would come in, consumers would engage, day one. And then there was about 60% to 70% of this audience that would just stop engaging.

Albert Thompson46:30The agent would follow up, follow up, follow up, follow up. About six to seven weeks, we saw 40% of this audience re-engage out of nowhere. And then in about two to three weeks, purchase. So there was a lag of about seven to eight weeks from four weeks of zero interaction, zero re-engagement, touched once, seven to eight weeks later, 40% of this group would actually then come back and they would buy. And when you were able to extract the data signals from within it, it came down to pretty much one thing. That particular audience was liquidity constraint. They, everything came back to, I'm getting a bonus.

Albert Thompson47:05I'm waiting on taxes. I'm waiting on money to come in. Yet they started shopping in Thanksgiving.

Brian Pasch47:10Yeah. What a great reminder that with AI, we don't have to worry about an agent getting tired or burnt out or quitting, especially with some of the third-party tools, even data from companies like Client Command that can see people when they're back out shopping, feeding that into an AI model. I mean, there's so much, so much in the future. It's going to be awesome. Albert.

Albert Thompson47:36Yes, sir.

Brian Pasch47:36CEO, IDPrivacy.ai. Albert, thank you so much for being on our podcast. And now that DMSC is over, we need to be providing clarity for dealers who want to upgrade their tech stack and leverage AI, but do it in a safe way. We'll be talking more of this with future shows and getting ready for MRC. Thanks, everyone, for watching. Albert, thanks for being on the show. And finish the month strong because we're going to help you sell more cars in a digital age.

Albert Thompson48:09Thanks.