"AI Can't See!" The AI Phase After Agentic That Will Reshape How Dealers Run Every Department (+ How to Use It) | AJ McGowan, VP of Research and Development at Reynolds & Reynolds
Car Dealership Guy Podcast
"AI Can't See!" The AI Phase After Agentic That Will Reshape How Dealers Run Every Department (+ How to Use It) | AJ McGowan, VP of Research and Development at Reynolds & ReynoldsCar Dealership Guy Podcast · Jun 25, 2026
Reynolds and Reynolds makes software that car dealerships use to run their day-to-day business. In this episode, it’s important because AI changes can roll out to many dealerships through their systems.
ChatGPT is an AI chatbot that can answer questions and write text. Dealers use tools like this to help draft messages faster, but it can still make mistakes, so companies add rules to keep it safe and accurate.
Sam Altman is a well-known tech leader associated with OpenAI. The episode mentions him to show that people initially doubted ChatGPT, but it ended up changing a lot of things.
Cognitive AI is the “next step” in AI beyond chatbots and basic automation. The goal is for it to understand what’s going on more deeply and help make better decisions in real workflows.
Agentic AI is AI that doesn’t just talk—it can actually do tasks. For example, it can take actions inside a dealer’s computer systems to update customers or trigger follow-ups automatically.
DMS means “dealer management system.” It’s the main software dealers use to run day-to-day operations like tracking customers, vehicles, and service work.
An omnichannel buying experience means you can start buying online and finish in the dealership (or vice versa) without everything feeling disconnected. The goal is a smooth, consistent journey across channels.
Carvana is a company that sells used cars with a big online-first focus. The point here is that dealers aren’t just trying to build one more app like that—they want a broader, multi-channel experience.
Machine learning is when a computer learns from lots of examples to make better guesses. Instead of being programmed with exact rules, it figures out patterns from data.
“Cognitive software” is a general term for AI-like programs that can understand information and help make decisions. Here it’s being used to describe the different stages of AI tools people are building.
An LLM is a large AI model trained on lots of text so it can understand and generate language. That’s what powers many “write/summarize/chat” AI tools.
The AMC Matador is a car that was made by AMC, a U.S. automaker, mostly in the 1970s. It was designed to be a comfortable, mid-size family-style vehicle. People may mention it because it’s a well-known older model from that time period.
Auto Vision is a software tool that helps dealerships estimate what a used car is worth. It gives dealers information they can use when deciding what price to put on a car.
It’s a computer system dealerships use to keep track of their cars and help decide what to list them for. Here, it’s meant to turn valuation data into practical pricing and buying decisions.
Bloomberg Terminal is a famous finance software used to look up market data and make decisions. The speaker is saying Auto Vision plays a similar role, but for used-car pricing and valuation.
Ray is another AI system the company is building. It’s meant to coordinate multiple tools and data sources so it can help dealers reach a specific decision goal.
Goal seek means the software tries different inputs until it hits a target result. Here, it’s described as searching through the dealership’s data to find the best answer for a pricing/decision goal.
CRM is the system dealerships use to keep track of customers and leads. It helps them see things like how many people are contacting them and what happens next.
To “reprice” a car means changing its asking price. The idea here is that software could automatically suggest price updates based on what the market is doing.
Term
VEC
VEC here is a pricing/valuation tool the dealership uses to estimate a car’s value. The segment’s point is that AI can show dealers what that valuation tool is doing so they can act on it.
Term
appraisal time
Appraisal time is when the dealership checks a car and decides what it’s worth. The segment says AI can help make those recommendations during that evaluation step.
The service drive is where cars come in to be serviced or repaired. The example here is AI spotting that certain jobs will need parts and ordering them ahead of time.
Inventory acquisition is how a dealership finds and buys cars to sell. The idea here is that more of that sourcing will become automated in the future.
It means the dealership’s different computer systems are connected so they can share the same customer and vehicle information. That way, the AI can make smarter suggestions because it’s not guessing with only one department’s data.
It means the dealership keeps information in a way that’s connected to the exact situation—like the specific car and the specific customer. Then the AI can use that connected info to suggest the right action, not a one-size-fits-all message.
An “oil change” is routine engine maintenance where used engine oil is replaced with fresh oil and the oil filter is typically serviced. In this segment it’s used as the trigger event for AI to recognize the customer’s vehicle and lease timing, then route them toward a sales opportunity.
A lease is like renting a car for a fixed period with monthly payments. When the lease is almost over, the customer is often thinking about what to do next—trade in or buy—so the AI can target the right moment.
A warranty is coverage that helps pay for certain repairs if something breaks. Dealerships often sell extra coverage, and the example is about AI using timing and vehicle info to suggest those add-ons.
A trade is when you bring your current car to the dealer and use it as part of the deal to get a different car. The AI is trying to spot when that moment is coming up.
A service rep is the person in the service department who talks to you about your car’s service. Here, the point is that AI should know how that customer usually communicates so the next step is smooth.
The idea is simple: AI can only use information it has access to. If the dealership doesn’t collect or connect the right customer and vehicle details, the AI can’t make smart, personalized suggestions.
CDP here means a system that gathers and organizes customer/lead info in one place. The discussion is about making sure the AI can “see” that customer data so it can help the dealer respond faster.
Term
agentic future
“Agentic” AI means the software can do tasks for you, not just talk. Here, they’re saying future AI will need access to the dealer’s data so it can take the right next steps.
Opportunity cost means the downside of not doing the better option. They’re saying if your dealer’s tools and data aren’t all together, you’ll lose out on AI-driven efficiency as time goes on.
“Consolidated” here means putting your important dealer systems and data into fewer places so everyone can use the same info. The argument is that AI works better when it can access everything together.
LIVE
in the first phase with generative. It's like we were taking a slice of lemon and dropping it into a drink. And in the second
phase with a genic, we're trying to squeeze that limit as hard as we can. In the third phase with cognitive, it truly should be
transformative. It's almost like we're making lemonade. We've got something that's completely net new.
AJ McGowan on the CDG podcast, our founder and residents acquired, company was acquired by Reynolds and Reynolds
about three years ago. That'll be a, that'll be an interesting story to talk to you about.
Yeah, yeah, it's been a fun ride so far.
So was this before chat GBT or after chat GBT?
So this one is when chat GBT was still pretty nasive. Not too long before we got acquired, we were definitely starting to
play with it and definitely saying this is not quite ready for primetime. I can't can't be trusted without a lot of guard
rails. But we were starting to really get a lot of interest in what we could do with all of us in the big picture for
sure. Talk about not listening to the haters. You know, there's this, there's this viral photo of Sam Altman, the founder
of or I don't found a co founder of open AI, where he's saying, Hey, we just launched this thing, chat GBT, I think you'll
like it, check it out. And there's this like famous comment underneath, which is like, dude, this is one of your worst
ideas yet. This is going to flop. And it has not. So it actually has revolutionized a lot of things. So yeah, well, I
think this will be a particularly fun conversation because while we've discussed about the trend of AI, you are
trailblazing this within one of the most impactful companies in our industry, which of course, Reynolds is
integrated into many thousands and thousands of dealerships across the country. Number one, so you have lots of impact
there. Number two, I heard you say a line which I haven't heard anyone say before about AI or just talk about in
this space, which is that we are now entering a new phase of AI called cognitive AI. We initially spoke about
generative generative AI, which every dealer here uses or has used, right? That's chat GBT. I mean, all these, you know,
LLMs, that's at least a form of it. We then started talking about a year and a half ago about agentic AI, where some
vendors that serve dealers started, you know, implementing agents, which actually do stuff, do things for you tasks, you
know, whether it's in your CRM, your DMS, right, things are actually happening automatically. Many dealers have adopted
that. And some have even done it themselves. And now there's this next iteration here, which you call cognitive, we'll talk
about what that means. Before we get into that, given you're such a, you know, you're an interesting technologist, I want to
start with like, just if we can hit on like a theme, like thematically, what is interesting to you nowadays? What's
exciting to you from your work in this industry?
What's so interesting to me about automotive specifically, especially when it comes to AI is the fact that in
automotive, it's really all about the customers. It's really all about the people and it's about that relationship that
dealers ultimately have with their customers. You know, I think in a lot of places, and in a lot of other industries, we're
seeing AI being, you know, looked at as a tool that can just sort of replace that relationship or replace those people or just
do a job. But in automotive, at the end of the day, that relationship with the customer is sort of the most important
thing. And so I think in automotive specifically, there's a unique opportunity to really leverage the promise of AI to help
employees be better at building and maintaining customer relationships. And then I think that that sort of unique, you
know, retail, if you will, the outlook on virtually every aspect of the business means that we can, we can leverage AI to
create new kinds of software. And, you know, think about software differently than we did in the past.
Dealers are not asking you, Hey, build me the next Carvana. They're saying, Hey, build me this like omnichannel buying
experience where I can run a tighter ship, have it be, you know, let my people focus on what they're great at. That's
semi accurate.
Yeah, for the most part, I think, you know, just tying that together with, you know, you mentioned a second ago, I've been
talking about this, this sort of cognitive software thing. When I think about sort of the phases that AI has gone
through, you know, there's sort of this like preliminary, you know, prehistoric AI, right, which we called machine
learning. And we've been working with for a long time, right? And then you have the the advent and by a long time, I'm
talking an AI term. So long time meaning, you know, the last like eight, 10 years, that's been something that you can do
interesting things with. You know, I think when most people start to really think about AI is your point, you know, chat
GPT was the invention of the transformer, which led to you, you know, the LLM, ultimately, and, and then, you know, chat
GPT and others that have come out and sort of brought that out to the masses. But when you think about what that phase
has really looked like, that was, you know, generative AI, right? It was, how do I summarize this email? Or how do I, you
know, come up with a great recipe? Or, you know, any of these other things where it was, you know, pretty bounded task that
you would just ask for. And then, you know, in the in the egenic realm, you know, as we talk about like the last like
year and a half that that term has really been popularized, where we're, we're fundamentally talking about is is
true like companion software, right? Like I like to say, sometimes generative is like an intern. And the egenic is like having
an employee, where it's a real companion, that we are seeing how to use the other software that we use, right? And so that, that
whole phase and everything that we're doing around that is super, super exciting. I think there's massive potential, you
know, we're, we're working on the egenic framework that we call Ray, the Reynolds, that's something that we've kind of been
spearheading in the industry. But what we talk about cognitive software, what I'm really talking about, the thing that gets me the
most excited right now is, okay, great, we take AI sort of to the point where, you know, we can use these LLMs to manipulate the
software that we're already using. But what does it look like when you start to think about how software works, period, with
a AI embedded directly into the software? One kind of fun definition that I've heard about cognitive software is, you know, the
way to think about it is, if you remove the AI, then the software doesn't work anymore. That's the jump from egenic to
cognitive, where it truly becomes not just a system that's built for agents to be able to manipulate it easily, but a system
where AI is deeply embedded inside of every transaction, every interaction, and the software itself, in a lot of ways melds with
the AI to become truly a learning system.
What dealership problems are you trying to solve with cognitive AI? Like if I'm today saying, hey, I am committing to a DMS and I
choose Reynolds, right? What am I expecting? This is, by the way, a question that dealers, I said, hey, I'm speaking with AJ from
Reynolds, who has, you know, questions and I have some auto vision stuff as well. But I think with regards to the overall bet on
Reynolds, what should I expect from your team as I think about cognitive AI? And at the end of the day, I want to know what
problems are you going to solve for me? What's that like the next phase there?
Sure. Yeah, I mean, so let me answer that question more broadly. So what can you expect for the bet for Reynolds? You know, the
first is continued work on all of the generative AI tools that we have. You know, we were first to market with a lot of those
tools made big bats years and years ago and have, you know, great services that will do all kinds of interesting cool things
that are embedded into the platform. The second piece is the agent piece where we're investing massively in our own hardware
inference, our own data centers, to actually run this AI, building out and training our own models like real,
agentic infrastructure, doing a lot of the hard side.
Can you touch on that for a second? Why are you investing in hardware? Reynolds is a software company, I believe, as far as I
know. Why are you investing in hardware? You don't hear that too often.
Yeah, so it's a great question. So when we look at where the component, you know, the place that AI has and sort of the future
of technology, you know, with the couple reasons for us to sort of own that infrastructure, one of those, you know, is
sort of the easy obvious one, which is we want this to be deeply embedded and endemic in all of our platforms, which means
that we need to control costs. We need to make sure that we can control what the cost of that inference is so that we can
continue to provide value to all of our customers at a great rate, right? So that's that's kind of number one. Same thing
around quality, like we don't want to be beholden on somebody else that might go down. Like if we're going to treat this not
just as sort of a novel addition to the software, but as something that's really core, we need to control that
infrastructure. And the third piece is privacy. I mean, a lot of the most interesting use cases for AI inference
involve data that, you know, is either on the line or that you definitely wouldn't want to hand out to a third party.
This episode is brought to you by Matador AI. A lot of dealerships are trying AI right now. Maybe you already have it
somewhere in your store in chat, email, CRM or follow up. But the real question is, is it actually helping you sell more
cars? That's where Matador AI comes in. Matador works, your leads follows up quickly, keeps customers engaged and connects
directly with your CRM. So your team spends less time chasing cold conversations and more time with shoppers who are ready
to move. And in a thin margin business that matters, every stale lead, every missed follow up, every slow response can
turn into lost revenue. So if you want to see what dealership trained AI can actually do, visit matador.ai slash CDG to book your
demo mentioned CDG and get your first month free, or just click the link in the show notes below. So the way I would explain
this to a dealer or simplify, I would just simply say that and it makes sense that every time we are using some, you
do some querying AI in some way. This is, you know, even today, like you, you know, I use certain tools, even the app we're
recording this podcast on right now. When this podcast ends, I will not automatically get, you know, like a transcripts and all
this, you know, like companion products or whatever you want to say, I have to click to get it. So I email the CEO, I said,
why do I have to click every time and wait 10 minutes? He's like, Well, because only, you know, a small subset of users do this.
And if we just ran it automatically for every single user, we would be spending a ton of money on, on inference, and they just, it
would be like, it wouldn't be economical. So I think what you're essentially saying is, yeah, like, for us to create the DMS of the
future, where a dealer has AI deeply embedded, or it's AI first, you have to control the cost in order to use that effectively,
which means going into hardware, which I think is fascinating.
So it's the cost, the stability and the security. Those are sort of the three main tenants. And, and to your point, like, we want it, we
want all of the software that we build going forward to be the best possible user experience. Now, we don't want to have to have
tradeoffs, like the one you just mentioned, you know, it's, we would love like, I'll pretend that I'm the, the product person now on
your, your recording software, you know, we'd love for it to the transcript to get recorded every single time and then send you an
email afterwards to say, here's how you did. And here's some of the key highlights and some cut scenes that you might want to pull
out. And by the way, if you want the short clips, then here's the links that you can send them to your social. We've already done that
work for you. Right. Well, that's, that's the kind of workflow that you can enable and that you can create these sort of
magical user experiences, when you can count on the AI to be just another component in the toolbox, as you build out software.
So AJ, there's Avery, there's Ray, of course, your claim to fame initially, auto vision. There's a lot happening at Reynolds. I guess you
wedged into this industry by way of inventory management. Right. That was where you started. You clearly saw that this one big fat line
item on the dealer's balance sheet called inventory could probably be handled a bit more efficiently. And so you built an inventory
management and, you know, analysis optimizes, I mean, many, many things on top of it, but pretty robust inventory suite. By the way, we
had some fun where we went on our CDG marketplace, which is where we have our dealer conversations. It's CDG app.ai. And I wrote auto
vision. I just wanted to see like, because I've never personally used auto vision. It was interesting, like people had some pretty good
feedback for you. So this is, you know, dealers can see this who are verified through our platform. But nonetheless, you're clearly,
clearly doing something right. And that realm, I would say that, you know, primarily as I'm reading this, people are very excited
or dealers are excited about they wrote more AI recommendations. So versus the market. So I would have to assume they're
referring to, you know, when it comes to like pricing a car acquiring a car, whether you're aware of it or not, that's
something that people find a lot of value in that you've built for the for for dealers the ability to, you know, as they say, use
raw market data and just create better recommendations for my inventory management through auto vision. But I don't want to
monopolize the conversation on that. I want to let you I want to let your brain wander here and say between every ray
auto vision, what is right now like, you know, where are you working to solve most problems for dealers? So when we talk about
auto vision, you know, the thing that really distinguished us in the marketplace and the problem that we set out to solve was how
do we get better at valuing cars? I mean, that was the fundamental like before, frankly, we even decided we wanted to build an
inventory management platform. The initial idea was, if we can get really good at valuing cars, there are cool, interesting use cases
for being able to do that. And so that was, you know, the kind of fundamental data science problem that we started on with
auto vision. What we realized, you know, over the course of time was to get it into dealers hands, because I think we all know
like, the last thing the dealer wants is one more tool, right, like one more tab, one more login, no one wants another tool.
Right. So we, you know, we can we almost got forced into building an inventory management platform, because otherwise dealers
weren't going to be able to access, you know, this mechanism that we built for valuing cars more accurately. And so we ended up building
the rest of the platform, sort of around that. So when you think about auto vision, it's really like a data science product. And then on top of that, we built
an inventory management platform to make it accessible. And what was interesting was, you know, right around the time of the acquisition, like
when Reynolds fired us, we sort of reached a point where I think the most generous way that I've heard this put is auto vision became a
Bloomberg terminal for used car dealers. And, you know, I think that the the generous part of that is that it acknowledges that we
surface a lot of data, historical trending analysis, and different ways to, you know, sort of look at what is the value relative to
wholesale or other people, etc, etc. But the reality was, and then, you know, to a certain extent, still is that in auto vision, we can
help you make hundreds of dollars more per copy, and then have seen that now as we've scaled out through Reynolds proven over and over
again. But it might take you five or 10 minutes to go through and look at all of the auto vision data and understand the decision that you
wanted to make. So one of the things that we did, you know, a few years ago was we built this, you know, it was it's funny because it's
before we started talking about agentic AI, right? So I like to say sometimes that Avery is kind of like the hipster agent,
because she was an agent before it was cool. But it was, you know, it was an attempt to basically say, how could we train AI to use all of
these tools that we have in auto vision to value cars, instead of just having like this arbitrary black box, or if like some weird model
that we trained and you know, some stuff, instead, we said, can we train the AI to use these tools and then evaluate it the same way that we
would train a user to use auto vision. And that, you know, ultimately became Avery. And is the thing that that you're talking about with
being able to recommend like, here's the number you should put on this car where you should reprise, etc, etc. But that that experience was sort of
the direct through line into Ray, where we said what's right. So Ray is the the genetic platform that we're working on right now. So it is an
agent that are a series of agents inside of this overall architecture, that we're teaching how to use all of the individual products now within the
Reynolds ecosystem, so that it can actually go, you know, go and goal seek against all of those individual data sources to try to answer users
questions and sort of, you know, objective number one.
Yeah, like AJ, if I start the month and I say, Hey, we need a bit higher volume this month, I want to blow out some of the older units.
And maybe, you know, the stuff that just came in, try to claw a bit more margin, and then it's going to find how to do this for me.
That's right. That's exactly right is to go and figure that out and say, Hey, you know, I need to look at some DMS data here, I need to find some
auto vision data here. Maybe I need to go and pull how we're doing out of focus and CRM. I need to look at, you know, what is lead volume look like on these
cars. And Ray will go and goal seek for you to come back and try to answer your question. And then ultimately, what we want to take that is Ray becomes the
organic infrastructure, then they can also goal seek to go and do something for you, right? That would be sort of the next step here is teaching her how to find data in the platform, teaching
her how to analyze that data appropriately to give you like real business value. And then ultimately, to let her take action on those insights to say, like in your example, you know, go find these cars.
Well, why don't you just let her go reprice them? And if you're happy with those results, maybe you let her reprice them every morning. So those types of things are what we're building the Ray infrastructure for.
Take me down to the dealership level. Right? What are what's the workflow today of a dealer? Has it changed meaningfully in the last 24 months? Like what does it look like today? How our dealer is really using this software?
It hasn't changed meaningfully yet, is the honest answer.
Dealer's workflow. You're talking when we talk about the generative stuff. So or in the in the course of like Avery, for instance, when you talk about the individual tools that are built into the platform, dealers workflows have changed massively. I mean, we have, you know, the
like, it would have taken them hours and to do it. But the reality is, because it was so time and time intensive and laborious, most of the time it just didn't happen. So for most dealers, that insight into what their VEC is doing is net new and has huge changes to their behavior.
Or if you look at auto vision and Avery being able to to recommend on cars, we can actually run that, you know, both at appraisal time. So it helps, you know, in their workflow when they're inspecting cars.
But it also can run every single night and look at all of their inventory and say above or below a threshold. Here's where you're at relative to where Ray says that you should be.
So instead of a used car manager now having to go through and hunt through, you know, 150 cars and say, how do I look relative to market? Why isn't this moving? Am I not getting enough leads on it?
You know, Avery looks at all of that stuff and just says, Hey, here's one I think you should go take a look at. Here's where I would recommend that you price it.
So all of those sort of discrete tools are changing dealer workflows every single day. The way that I think about Ray is now, how do you wire together those tools to then produce even bigger outcomes, right?
To produce, you know, both strategic insight outcomes on a daily basis, you know, things like I'd like, I'd like for Ray to be able to say things like, Hey, you have six cars coming into the service drive today and we're missing parts on two of them.
Would you like to go ahead and order those parts in?
And those are those are the types of things that we would expect in this next wave as a genetic goes beyond just where it's at now, which is go find me this thing and goes to being proactive with giving you information and then ultimately being able to take action on that for here.
So based on the last example, you just provided AJ, where is a dealer to find an edge nowadays? Like what has, you know, what has staying power, right?
Because if, if you're saying that the software is increasingly doing all these tasks and we'll do more of them, right? Parts ordering inventory management, I'm sure, I mean, you already have big companies doing automated inventory acquisition to a certain extent, not that dealers can't,
but I'm sure that's going to proliferate over time. So like, where's the edge for the dealer? If I'm saying, okay, what can I invest in, whether it be my time, energy, capital to create differentiation over time?
Yeah, you know, it was like, you, AJ, you used to have like, Oh, like, you know, he's been on the block for 40 years, he knows all the prices and he like, he knows the market, but you cannot compete with an LLM and what you're building working on.
You just cannot over, you just can't do that. So that is not going to be an edge in the future. Where, where is the edge for me to invest in today as a dealer?
Well, so I think, surprisingly, it's still there. It's still in your people. The, the ultimate ceiling here, right? So we talk like everything that's happening in Agenic right now is super exciting. And there's a lot more work before we're going to hit bump our heads on the ceiling.
But the ceiling here is going to be that running these agents, they only have contacts for what's happening right now, unless you go and drag in a bunch of logs and analysis and other things from other places and give them this kind of massive contacts to say, make a decision right now about what I should do.
And that's, that's the money right there, by the way, that's your investing into hardware to be able to do that.
That's right. That's exactly right. And that's, that is, you know, being able to do that more and more effectively is going to make these agents more and more effective. But there's a point where we're going to hit our heads on, that's just not official. That's not, it's not efficient.
And it also still requires that a lot of very specific contextual data goes in and it's wasteful, right? Because in a lot of cases, contacts use for a particular question at a given time, given other contacts, some things are required and some things aren't.
And so where we're going to see, where we're going to see the sort of ceiling there is that ability to understand what is the right context at the right time. And that requires that we go and say, like, go look at this, this.
And so, you know, the hardware, the hardware investment, as you said a second ago, is sort of critical to that and making that scale up and scale up well. But we're still going to bump our head on the ceiling because it's only going to get better and better. We're going to have more and more data.
So when we start talking about that third wave that I see in the future, which is cognitive software, part of the idea there is that the software itself retains a memory and understands the context that's happening at every given moment because the AI is constantly looking at all the factors as everything is changing in the system.
And so you end up with like a lot of people in AI more generally call this concept like institutional memory, where, you know, it's baking the institutional memory into the software.
But the problem is like one way to look at that is that gives you then, like I said earlier, it's like the intern and the employee with like cognitive software to be like the veteran.
You know, it's the veteran that's been there for 30 years, like you were just talking about.
And the reality is that, yes, it can make that new hire have the kind of insights, because it has access to that context all the time that a veteran would.
But investing in that new hire so that they can have that relationship with the customer and they know what to do with that data still is the most important thing.
And that's that's going all the way back to the beginning of our conversation.
Why I'm so excited about doing this automotive is that that that people part right where the people meet the customer is something that I don't think you can ever get rid of inside of automotive.
It's something that is, you know, the core when you strip away all the operational complexity, that's the core of the dealership's business is that trust and that relationship with that customer.
And so I think on the system side of what what dealers can be thinking about investing in is trying to get their data onto common platforms so that AI can have contextual access to it.
And from a business standpoint, I think it's going to be the same answer that it's always been, which is that the dealership is really about the people that work there.
And in a cognitive software world, those people continue to feed the software in this virtuous loop that gets you better and than the software decisions tighten in around the dealership and the people and the market and the cars and the customers.
So you're getting higher and higher quality results from it.
But investing in the people is the thing that keeps that thing working.
This episode is brought to you by Zurich.
Dealers, if you want more consistency and better performance out of your FNI teams, you need more than guesswork.
You need real insight.
Zurich Advisor IQ uses AI powered coaching and actual deal data to show you exactly what's happening in your FNI process, what's working, where you're losing opportunities and how to improve.
With personalized feedback and real world coaching, your team can build stronger customer conversations and deliver better results deal after deal, ready to see what's really driving your numbers.
Visit Zurich NA dot com slash Zurich Advisor IQ or click the link in the show notes below.
So two things you just said there, and I want to touch on the first one you mentioned context or like having the systems talk.
I want to understand what you mean by that or how a dealer can do that as best as possible and what's in it for them.
As a sidebar, we love to talk about like picks and shovels on this podcast and I can't help but wonder as you talk about this, you know, where people will be the where the dealership excels because that's essentially the last part that's not commoditized over a long enough time horizon.
It makes you wonder like what the opportunities that lie ahead of us in stuff that is like as sounds simple, but like training, right, like investing in culture, like these, these type of activities that historically were, you know, not anything sexy or fancy, you know, anything special.
But when you think about the future and if every dealer ship is increasingly using, you know, software, which is commoditized, how do you stand out?
You stand out by having the best people and how do you have the best people and getting there.
So anyways, this is just where my mind wanders.
Going back to the first point you made, you said contextual databases, I believe you're talking about having as a dealer, invest in having your systems work with each other, talk with each other.
Can you tell us what do you mean by that?
Yeah, what I mean by that is that the more the more advanced this AI becomes, the more that you're going to want your systems to be able to talk to each other in a deep way.
Let me give you an example.
We have somebody that pulls up on a service drive and they're just there for an oil change and they've been sitting on the car for, you know, three years.
What we want the AI to notice and be able to surface to the user is this particular car.
They've got two payments left on their lease.
And by the way, this is a great car for us in the market right now.
And by the way, we usually sell, you know, $3,000 in warranty and, you know, those types of things on it, as well as, you know, two grand on the front end for these types of cars.
This would be a great trade for us.
And we've got one that looks really good right now.
And she is a repeat customer that always talks back to her service rep right away and she prefers to be texted and so on and so forth.
Right? The net out of which would be, hey, this is a car we probably want to try to require.
Let's send her a text and say we may be able to give her a great offer and have her walk onto the sales floor to go talk to somebody.
That type of insight, well, it seems, you know, straightforward actually requires a lot of different pieces of data that come from different parts of the system.
AI can't listen to what it can't hear.
And, you know, that is like 100% the case when we talk about these systems because it doesn't have access to the contacts.
If all it knows is, hey, we grossed well on these kinds of cars and there's somebody in the service drive.
Yeah, you know that you're interested in it.
Maybe you text your used car manager and say, hey, there's somebody out there that you can go take a look at.
But you don't really have the full context to be able to say, no, this is a hot opportunity for us.
Here's why we need to get a hold of this person.
And if they don't come back in and go get a hold and send somebody out to talk to them, this is their preferred communication method.
Let's go.
Is that my job or is that my CDP's job or is that my my DMS's job?
Like, whose job is it to do what you just said to put all the information together?
Who should I be leaning on?
Like, whose best position to do that successfully?
I think, you know, quite frankly, we're best positions to do that effectively.
You know, that's that's who's we like it.
Is it like you're saying like you being a platform, which will your DMS and CDP.
So you're kind of NSCR and many other things under one roof.
When you think about an egenic future, having all of your tools under one roof is going to become much more important than it was in the past.
You know, I think that's that's exactly where I'm going to.
Yeah, it's hard for me to say because it sounds like a commercial break.
Like, hey, you're going to want all of your tools to be with Reynolds.
But the reality is it's just the truth.
I mean, you're going to want all of your data to live in one place that your AI can see it and you're going to want it to be able to see it in real time.
Yeah, I think and I could like to like, un-commercialize that.
I think the point you're making is you're going to want to have all your stuff consolidated and that increasingly the opportunity cost of not having all your software and data together will become higher over time.
Which is to say that we'll probably see just more consolidation in the industry.
And because dealers are are not dumb, they're going to notice that they're going to see they're getting better results, better profitability, better customer experience as they consolidate all their tools under one provider.
And there's not many players out there that can do, you know, most things under one roof.
But yes, I agree with that.
I think that it just it makes sense.
I would say historically and the reason I even personally was very keen on going to moving to a platform, which was, you know, everything consolidated under one roof was simply because, you know, from a training perspective, like it was just too difficult when you have to switch between platforms and click on all these different tabs.
But I think I think the era we're going to we're heading to now is to your point, it's the opportunity cost will just continue rising by not having things combined and being able to leverage the best possible insights and intelligence.
One of the interesting things about AI, you know, going back 15, 16 years, one of my really, really good friends is is a pretty, pretty well known AI, you know, machine learning engineer, just sort of in the circles.
And he taught me most of what I initially knew about machine learning 15, 20 years ago, and I've kind of followed through from there.
But one of the things that he said to me when we were going through it was, you know, AJ, one thing to bear in mind is until the moment of singularity, if a person can't do it, the AI can't do it.
If you could not teach a person to do this, then the AI is not going to be able to do it.
And, you know, this is way before LLMs or anything like that, but it's sort of borne out to be true over the course of all the iterations that we've gone.
And then it's still true is that the very best AI is only going to be able to do as well as the very best person.
And so if you'd expect for a person to have the problem, you should expect for the AI to have the problem.
I like it.
So basically, here's my takeaways from that, from that part of conversation, right?
The opportunity cost of consolidation will rise.
It's my best interest to work with a consolidator in that sense who can leverage AI the best.
The cost is obviously it's going to be increasingly very meaningful.
And I think inquiring into like the hardware investments that my partners are making so that I can have the best software and use just the best quality
product out there.
And lastly, that cognitive stage, which you mentioned, which is next after agentic, which just takes it to, like you said, it's AI first, without the AI can't even function.
That's how I know I'm really solving my problems as efficiently as possible.
And, you know, squeezing all squeezing the lemon just to the fullest extent.
I'd love to use your lemon metaphor down because I've never thought about it that way.
Yeah, it's almost like in the first phase with generative, it's like we were taking a slice of lemon and dropping it into a drink.
And in the second phase with a genetic, we're trying to squeeze that limit as hard as we can.
Yeah, in the third phase with cognitive, it truly should be transformative.
It's almost like we're making lemonade.
We've got something that's completely net new.
And that's, you know, kind of how I look at the future that we're trying to improve and until we've squeezed all the juice out that we can.
And then it's like, OK, well, now what can we do with all this juice?
Amazing. Look at that. We're even coming up with analogies.
AJ is going to send me some royalties for that one, folks.
AJ, any closing thoughts?
I think just to close out, it's a really exciting time.
I mean, this is, as I hope has come across in our conversation, I think automotive is a place where we can be at the absolute bleeding edge of this technology and get the most squeeze out of it.
And so it's really exciting for me to be in the middle of it, working with customers, getting their feedback as we're building these tools in real time.
And quite frankly, as we're all figuring out together as a species, you know, what place AI has in our lives that we get to have a front row seat for what place does AI have in our dealerships?
And, you know, that's a it's a really fun thing.
And it's really fun to be at Reynolds and Reynolds where, you know, we've committed resources, time, attention, you know, the whole thing's being all in on AI and get to lead that charge.
So we're really, really excited about, you know, what the next few years holds.
Amazingly well said. AJ McGowan, founder of AutoVision, acquired by Reynolds and Reynolds and now a senior leader at Reynolds.
AJ, thanks so much for joining us on the podcast.
Thanks for having me.
All right, hope you enjoyed that episode.
Please give the podcast a rating, consider subscribing to the show and check the show notes for links to the sponsors of today's episode, Matador AI Zurich, and of course, Reynolds and Reynolds.
Thanks for tuning in and I'll see you guys next time.
About this episode
The conversation maps AI’s evolution for dealerships—generative to agentic to cognitive—and explains what each phase changes in day-to-day operations. Reynolds & Reynolds frames agentic AI as software that “actually do[es] stuff” inside CRM/DMS, while cognitive software embeds AI so “if you remove the AI, then the software doesn't work anymore.” The guest stresses that AI can’t “listen to what it can't hear,” so success depends on consolidated, real-time data, controlled inference costs, and privacy safeguards—so teams can focus on trust and relationships.
AJ McGowan, VP of Research and Development at Reynolds & Reynolds.
He breaks down why fragmented software platforms are quietly limiting what AI can see and do, uses a service drive scenario to show how a fully connected system could surface a high-value trade opportunity in real time, and explains why Reynolds is investing in its own hardware infrastructure to make that scale.
Topics:
03:25 Why Auto AI Is Different.
06:45 Real AI Vs. Gimmicks.
08:45 Why Reynolds Is Buying Hardware.
13:15 The Bloomberg Terminal For Used Cars.
17:30 What Ray Actually Does.
22:50 Where Dealers Still Find An Edge.
28:35 Why Consolidation Is Now Critical.
This episode is brought to you by:
1. Matador AI - 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 @ here.
2. Zurich - Zurich Advisor IQ is Zurich’s AI-driven training and coaching platform built to help F&I teams perform more consistently and sell more effectively — using real transaction data, not theory. By analyzing actual F&I transactions, Zurich Advisor IQ helps identify behaviors and trends influencing results, delivers actionable insights and roleplay scenarios, and gives dealership leaders visibility into performance across managers, stores and rooftops. Connect with your Zurich representative to request a demo and see how Zurich Advisor IQ can help turn F&I insight into stronger dealership performance. Discover more @ here.
3. Reynolds and Reynolds - Turn cars faster and increase profit with AutoVision, an end-to-end inventory management suite that optimizes every step of the used vehicle lifecycle. From acquisition to sale, AutoVision gives you a clear way to manage inventory. Visit AutoVision.com for more information.
Check out Car Dealership Guy’s stuff:
For dealers:
CDG Circles ➤ https://cdgcircles.com/
Industry job board ➤ http://jobs.dealershipguy.com
Dealership recruiting ➤ http://www.cdgrecruiting.com
Fix your dealership’s social media ➤ http://www.trynomad.co
Request to be a podcast guest ➤ http://www.cdgguest.com
For industry vendors:
Advertise with Car Dealership Guy ➤ http://www.cdgpartner.com
Industry job board ➤ http://jobs.dealershipguy.com
Request to be a podcast guest ➤ http://www.cdgguest.com
Car Dealership Guy Socials:
X ➤ x.com/GuyDealership
Instagram ➤ instagram.com/cardealershipguy/
TikTok ➤ tiktok.com/@guydealership
LinkedIn ➤ linkedin.com/company/cardealershipguy
Threads ➤ threads.net/@cardealershipguy
Facebook ➤ facebook.com/profile.php?id=100077402857683
Everything else ➤ dealershipguy.com
