Written by Alysha Webb
AI continues to dominate conversations across the automotive industry, with dealerships exploring how it can improve efficiency, streamline operations, and unlock value from the vast amounts of data it can absorb. While much of the attention has focused on marketing and customer engagement, the potential impact of AI reaches far deeper into dealership operations.
To better understand where the technology is headed, R&Co spoke with Todd Smith, CEO of QoreAI. While much of today’s AI discussion focuses on applications and automation, Smith argues that successful AI adoption begins with something more fundamental: data ownership and organization. QoreAI was built to help dealerships centralize and leverage data that often resides across multiple platforms, creating a foundation for more effective use of AI.
In this conversation, Smith shares his perspective on the role of data in AI adoption, how dealerships are using AI to improve efficiency, and why he believes the technology will fundamentally change how dealerships are organized and operated in the years ahead. This is an edited version of that conversation.
Building a Data Foundation for AI
R&Co: What does QoreAI do for dealerships? What is it replacing that did not use AI and that now dealerships can utilize AI to handle or improve on?
Smith: I felt that for all dealerships, all of their data is being held disparately in systems – DMS systems, CRM systems, things they don’t own or have full control over. So I said, well, if anyone’s going to leverage AI, you have to do it from a data layer, and then you build on top of it. Dealers can’t do that, they don’t own their data layer. They are renters, in a fiefdom basically.
Now, in the world of AI, the data is being put in those tools and the derivative data being created out of those systems is incredibly valuable and you don’t want the underlying vendors to be the real gainers of that. You want the dealership to use that data to continue to grow and have improved capabilities. So, I felt the first step was helping dealers get all of their data out of these disparate systems and into a system they own. So, we set up every dealer in their own system of AWS (Amazon Web Service). The app layer is shared, but the dealer data itself is separate and owned by the dealer.
And then we take all the data out of their systems. We handle the data hygiene, cleaning, filling in missing data points, ensuring everything is mapped correctly, vectorizing the data, doing what I consider very grunt tactical work that has nothing to do with AI. It’s just all block, tackle Machine Language work and database work.
Expanding Insights Through Data Partnerships
R&Co: So where does the true AI come in?
Smith: Once we do all that, now you have clean, organized data that’s constantly being updated, coming through CRM and DMS, everything is correlating. We said, okay, dealer data is very valuable because its transactional in nature versus someone looking at a website. We know someone wrote a check, right? They bought a car, sold you a car, or serviced a vehicle with you. So, we said, okay, that’s good data, but it’s not enough. So, we partnered with Equifax and information on approximately 318 million Americans is updated and made available to us every 90 days. (Equifax has individual’s credit account, collection accounts, public information related to bankruptcies, foreclosures and tax liens, etc.), which gives us much deeper intelligence around the psychodynamic understanding of the individual, beyond the transactional data coming from the dealership.
R&Co: How is that information used?
Smith: We know everything like all the addresses you lived at, previous phone numbers, email addresses, your income, your financial durability. We have it all, everything except your FICO score. That helps us build much better engines, like algorithms to be able to help and analyze customers beyond the limited data that comes out of the dealership.
Every dealer should have a data foundation, so that’s what this is. On top of that, we run an AI intelligence layer, so for every dealer group we build a RAG (Retrieval, Augment, Generation). In simple terms, it’s a filing cabinet. So, the goal with this filing cabinet is every dealer operates uniquely in their own space.
R&Co: And what is that used for?
Smith: All of the information that makes the dealer unique – some are on price, some negotiate, etc. – is put into the RAG. Then we have a deep partnership with Anthropic’s Claude. We use Claude in the background to build an intelligence layer over all of that that the dealer can use to get faster answers to their questions instead of slogging through reports to try to analyze things. AI cuts through that noise very fast. So for us, we want to help our dealers and groups get their answers faster.
That’s one side. The other side is we’re living in this new agentic tooling world. We’re not going to build those tools. We’re data infrastructure for dealers and groups, so we’ve built well over 200 connectors at this point so dealers can connect tooling, so if they want to control more of their marketing or social marketing or whatever, they can connect those tools directly to their data.
Improving Efficiency and Reducing Costs
R&Co: How many dealers are using your operating system?
Smith: We have a few hundred dealers that represent maybe five, six hundred franchises.
R&Co: What are some vendor tools or other functions that QoreAI has replaced?
Smith: A lot of dealers use third-party marketing services to do email. Now a lot of them just do it themselves, so that’s been a good cost reduction.
Like, one of our groups is saving $1.3 million a year just in email marketing. That’s easy, low-hanging fruit.
We just had a couple of our groups that had outside reporting tools they were paying a couple of hundred thousand, or a couple of thousand dollars a month per store. All that’s gone. They’re just able to ask questions of their data and get better answers framed, you know, like you have a reporting tool. It’s very structured. The report pulls from these tables and fields and generates this very specific outcome. With AI, you can ask a question and then continue asking deeper questions and as long as it has access to that data, you can get a much more comprehensive understanding around it rather than, what I would call, a flat line answer.
R&Co: What’s an example?
Smith: Well, first off, I go why? And what happened? We can go in and explain that a High Country Silverado takes 47, on average, vehicle detail page views to transact. So, we tag and monitor all of that data so we see, you know, oh, this one’s got 20-some to go statistically, on average even, to get the vehicle sold. There are layers of data, and our goal is to help dealers peel those layers back, so dealers make better business decisions.
R&Co: So, is this mainly a marketing tool?
Smith: Mostly operational. I think marketing is, you know, to us it’s a novelty. I think a lot of marketing will be completely agentically done. I don’t see humans really doing a lot of marketing for very long, unfortunately. I see it just being replaced completely by AI and agentic tooling.
So, we’ve leaned into far more structural operations of the dealerships by looking at inventory, reconciliation, and warranty submission. Just looking at sales and service defection, looking at a deeper insight into how the store is operating. All the way down to payroll.
R&Co: And so, does this just make your dealership a lot more efficient?
Smith: Yes, and in the end, that’s the right word. The real end game is to be more efficient than your competitors. Because if you’re more efficient then most likely you are more profitable. And if you’re more profitable, you can beat every one of your competitors.
Reimagining the Dealership Operating Model
R&Co: Where will AI take the dealership business? How do you see this advancing?
Smith: Fast. Really, really fast. And I think we’re going to see a completely different organizational design and structure over the next three years, five years at the absolute max. And I don’t think it will take that long. What I mean by that, I think, is you’re just going to have to almost sit down and reimagine how a dealership runs from a completely different lens. Right now, our entire organizational structure of a dealership has not altered much in 50, 60, or 70 years. We’ve seen nuanced changes. F&I came into play. Used cars became kind of a thing and BDC, all because of the internet. But I think this is a massive change to how we operate because many tasks that are primarily computer-based, whether they’re text, video, or data, can increasingly be handled by agentic AI. So, when you think about that, your entire dealership operation will be remapped.
R&Co: How do you mean?
Smith: Well, maybe a warranty clerk is out sick, or on vacation. What happens? The deals stack up on that desk because no one else knows how to submit warranty claims to the OEM.
The reality is, you could train an AI agent to do that automatically. I think people (working in a dealership) will have to adapt to new workflows that are more human-centric, meaning things that you still need a human to do, like a physical conversation with customers, hop in a car and go for a ride with them, tell them to take that payment, even if it’s $10 more than they wanted. Those types of things, I think humans are really built for because it’s a trust layer.
Now, a lot of back-end infrastructure of a dealership, whether it’s accounting, reporting, or other computer-based tasks, much of that work could become automated. I think the smartest dealers will lean heavily into it and automate.
Training AI Like a Digital Employee
Smith: AI is a different beast. It’s not like SaaS. It’s completely different; it’s like hiring a digital employee. So, we need to hire them, train them, give them feedback and coach them, continuously coach them.
R&Co: Because AI learns through iteration?
Smith: Yes, back and forth. It’s like a conversation. I have conversations with my agentic agents all day, every day. Sometimes you have to rein them in and say, “Hey, you’re doing something I didn’t ask you to do,” which they will do, if they think it’s helpful. So a lot has to be how do we frame it a little bit more narrow, yeah, but it’s fun.
Looking Ahead
While the automotive industry continues to debate the pace and scope of AI adoption, one theme emerged clearly from our conversation with Smith: the quality and accessibility of dealership data will play a significant role in determining how much value organizations ultimately realize from these technologies.
Whether AI is used to improve reporting, streamline workflows, support marketing efforts, or automate administrative tasks, dealers will need reliable data and thoughtful implementation strategies to achieve meaningful results. As the technology continues to evolve, dealership leaders who understand both the opportunities and limitations of AI will be best positioned to make informed decisions about where it can create value within their organizations.