Welcome to another edition of the Car Dealership Guy Industry Spotlight Podcast Recap—a rundown of key lessons from top operators, founders, and execs shaping the future of auto retail.
Today’s guests are Mike Martucci, GM at Jim Ellis Kia, and Kevin Esmezyan, Chief Product Officer at Matador AI.
Together, they discuss how conversational AI is turning a three-person sales team handling 700 leads a month into a machine that doubles close rates and books service appointments at nearly 90%, all while most dealers are still missing after-hours calls and watching customers drive to the competition.
Kia retention starts low and has to be built from scratch
Coming from a VW store where the customer base is naturally more dealer-loyal, the shift to Kia means earning every repeat visit rather than assuming it.
"It's really a retention piece from the dealer body to get that client back for the most important aspect of their car ownership, which is the repetitive maintenance.”
— Mike
Mike says the expectation gap is the core challenge, and that closing it requires education, consistent process, and enough trust-building visits that the customer stops comparing on price alone.
Dealership oil changes are often cheaper than independents.
The perception that franchise dealer service is expensive isn't always accurate, and Kia of Kennesaw is pricing aggressively to prove it.
"We are hyper price competitive. We research this just as much as we research a used-car price before it goes online. And at the end of the day, nobody's getting rich doing oil changes, even the independents. That's something we look at as a retention tool and basically a break-even piece."
— Mike
The specific education win is explaining that what independents advertise "starting at" is almost never full synthetic, and that the dealership's single starting price covers all product without the fine print.
Getting retention from the 30s to 50% started with including the first service at no charge.
Service retention was poor when the store started, but they found a fix in removing the friction on that first visit entirely.
"We were in the 30s. It was poor. This year we're going to hover around the 50% mark. A lot of that is is including that first service at no charge. That's a big one, obviously, getting them back."
— Mike
On higher-ticket products, the first two services are included because a customer who returns twice is far more likely to become a long-term relationship than one who returns once.
Context is what separates a good AI system from a frustrating one.
A lot of AI tools in dealerships operate on rigid logic like… if the customer says this, do that. The problem is that customers don't follow scripts, and a system that can't read context will misfire constantly.
"There are a lot of AI tools out there that are if-then. If the customer says this, do that. If the customer says that, do this. The challenge with those tools is they're not smart, and they don't understand the context of the conversation."
— Mike
Put simply: A system that escalates when uncertain and learns from the correction gets better over time, while one that just fires the wrong pre-programmed response creates the exact customer experience dealers are trying to avoid.
Presented by:
MatadorAi - Most dealerships are losing leads in the follow-up. Matador AI fixes that. It's not generic automation. It's dealership-trained AI, built on millions of real conversations and optimized around what converts. In a thin-margin business, every conversation matters.
Book your demo today and get your first month free @ http://matador.ai/cdg
Doubling close rate came from compounding small adjustments over time.
Trade-in valuations were converting below average. The fix was pulling value with a license plate instead of asking for a VIN first.
"There are small adjustments like this that are done over time and the compounding effect of that is, like you said, you double your your closing rate, you double your sales month-over-month. It's not a single switch, but as you activate these various automated messages, AI working in after hours, answering messages, answering calls, responding to customers, booking appointments, it just works."
— Kevin
The key to trusting AI with customers is being able to train it to behave exactly like you would.
The biggest concern about deploying conversational AI with real customers is that it says something you wouldn't say. The answer to that concern is access to the back end.
"The biggest thing for me with Matador is the access to the back-end to train that system to be like me. I can tell it exactly what I would like it to do, and I can tell it what to completely not do. And in the off chance that it's confused and it's not sure, it serves it up to my management team to confirm that it's right. And if it's not right, it educates itself on being right the next time."
— Mike
Over time, the alerts decrease because the system has been trained enough that it's right nine times out of ten without needing to ask.
A good GM in 2026 now has to know how to develop AI, not just manage people.
The job description for general managers has expanded, and the dealers who recognize that early have an advantage over those still treating AI as an IT purchase.
"It occurs to me that the role of a GM in 2026 now includes, for the first time ever, leading, managing, and developing AI. A good general manager in 2026 needs to understand it, needs to know how to buy it, needs to know how to articulate ROI, and then needs to know how to develop it over time. Fair?"
— Sam
Mike’s response: "Fair. You got to play the game where the ball's at."
Service books at nearly 90% when AI catches the calls that would otherwise go unanswered.
A service customer calling at 6:30 p.m. who can't get through doesn't usually wait to hear back; they usually just book somewhere else. The math on how many of those calls happen after hours is what makes the case for AI in fixed ops.
"We see numbers close to 90% booking rates for service customers. It's a different conversation. For service customers, they're looking to schedule service for their car. Whether this dealer does it or the dealer down the road does it, for them there really isn't much of a difference. Ultimately, comes down to two things: price and experience."
— Kevin
According to CDK, 25% of all service calls get put on hold, with an average hold time of 9 minutes, which means the missed-call problem is substantially larger than most service directors realize.
AI should be a pipeline for human connection, not a replacement for it.
The full-lifecycle vision is sales to service to trade-in and back again, with AI managing the handoffs so that every customer touchpoint leads to a human conversation at the right moment.
"Matador as a whole was built to maintain and help that customer throughout their journey from the starting point of a cold lead all the way down to a customer who's been servicing with you for 3 years and who won't buy a car anywhere else."
— Kevin
The bigger framing: AI handles the volume, humans handle the relationship, and the store that gets both right is the one the customer doesn't leave.
Dealers who treat AI as a switch they can just turn on will be disappointed.
The stores that get results from conversational AI are the ones that go through the onboarding work by training the system on their specific policies, pricing, and escalation logic before going live at scale.
"For dealers, I feel like the pitfalls that they can fall into is they'll see AI and they'll think of it as kind of a solution to all of their problems and they just want to activate it, and activate as much as they can as quickly as they can. And so it really takes a partner to work with the dealer to say, 'Hey, let's look at your business. Let's see where you're dropping the ball, where we can help the most right now. And then let's put a plan to scale you up over the next 3 months.'"
— Kevin
Mike's closing thought on implementation: "You've got to stay engaged as you go throughout the process. Otherwise, it just flat out won't work. You can't abdicate any of that.
