By Michael Quigley, President, Chief Strategy Officer & Co-Founder, Impel
I’ve been tracing how new ways of buying take hold in an industry, and a pattern keeps surfacing that I think we’re underrating in ours. A machine never steps in through the grand entrance. It comes in through the side door, the smallest and most routine transaction it can be trusted with, and it earns its way up from there. The ATM began with a twenty-dollar withdrawal. The banking that now lives on your phone came later, once the machine had earned the simple job first. Self-checkout started with “10 Items or Less”, not a full cart. Each time, the machine takes the easy, repeatable job first, and moves toward the hard one only once it has proven it can handle the small one. I’d half-forgotten how consistent that pattern was until I went looking for it.
AI agents are going to follow the same path into the car business. Agentic commerce is inevitably coming for the automotive industry, and once you see the pattern, it tells you not only where agents will enter but something more fundamental that much of the conversation is missing. The change isn’t really about how the deal gets done. It’s about who gets to be in it.
A while back, in Agent vs. Agent, I argued that the buyer would arrive at the deal with an AI of their own. That’s no longer a forecast. Shoppers already lean on agents to research and compare, and the first tools that negotiate on a buyer’s behalf are here as well.
This isn’t the last decade’s story of car buying moving online. That shift moved the shopper from the lot to the screen, but a person was still doing the shopping. What’s happening now is different in kind: the shopping itself gets handed to a machine. And a machine doesn’t just look, it reads, compares, and acts. Which raises a sharper question than whether your team can out-sell it across the desk. It’s whether the agent can read your store at all.
The agent decides who’s on the list
When an agent does the shopping, it does something a human shopper rarely did so cleanly. It builds a shortlist. The store it can find, understand, and actually transact with earns a place on that list. The store it can’t read never makes the list, and never even knows it was in the running.
For as long as any of us have been in this business, a lost sale was a fair fight lost, to a sharper price, a faster follow-up, or a better experience down the road. The agent era adds a round that happens before that fight ever begins. The store that isn’t legible to the agent doesn’t lose on the merits. It’s never considered at all. You can be the finest closer in your market and still never reach the table.
Read the same shift the other way, though, and it becomes the widest opening in years. A store an agent can read cleanly and quickly wins business it would never otherwise have seen, because the agent delivered the buyer to its door. The retail math hasn’t changed. The way in has.
Agentic commerce in automotive is inevitable
None of this is waiting on permission to begin. In September of 2025, OpenAI and Stripe published an open standard for agent-driven transactions. Visa introduced Intelligent Commerce and Mastercard introduced Agent Pay, and the largest payment networks and AI labs alike are maneuvering to occupy the middle of these transactions. Companies don’t invest at that scale for a feature. They invest for a position. McKinsey projects that agentic commerce could move $3 trillion to $5 trillion of global consumer spending by 2030, and the earliest dollars are already moving, through the simplest purchases first. The routine, low-stakes transactions are going agentic well ahead of the complicated ones, which is precisely why the first agentic transactions in your store won’t look like the ones everyone’s bracing for.
In a dealership, the simplest purchase is service, not sales. Booking a visit, approving routine work, settling a service bill is smaller and more repeatable than a financed purchase with a trade and incentives layered in. So the agent may well show up at your service drive before it ever shows up for a car. Which is exactly why the layer that makes a store legible can’t stop at the showroom. It has to run the length of the relationship, sales and service alike.
The car is the hard end, and where legibility matters most
The vehicle purchase sits at the top of that climb, not at the base of it. To carry a car deal to completion, an agent has to move through real pricing, live incentives, a trade valuation, a payment the customer can genuinely qualify for, and the history your store already has with that person. In our own data, drawn from nearly 50 million shopper conversations, financing questions alone account for close to 15 percent of everything buyers ask. No general-purpose agent can hold that much industry-specific complexity on its own. It needs a layer built for the work, an AI Operating System that sits between the agent and your store and represents the whole of it, accurately and in the moment. Not a rack of disconnected point solutions, but one connected system able to speak for the entire store at once.
That layer doesn’t push the dealer aside. It’s what renders the dealer legible in the first place, sets you in front of the agent, and keeps you at the center of the relationship, where the business has always been won. Here is the part worth sitting with for a moment. The deciding factor is not the size of the transaction. It’s whether a machine can read your store, and that holds as true for the small transactions arriving now as for the large one still on its way.
What it takes to be legible
Being legible to an agent is plainer than it sounds. It means your inventory, your pricing, your incentives, and your customers’ histories are connected, current, and readable by a machine, rather than scattered across systems that don’t speak to one another or held in the memory of a single advisor. When an agent asks the questions your customers have always asked, what the real payment is, what the trade is worth, whether a given rebate applies, your store either has an accurate answer at that moment or it doesn’t.
That is the test, and any dealer can begin grading against it today. A handful of honest questions get you most of the way there. Is your data consistent everywhere an agent might encounter it, or does it drift between your website, your feeds, and your CRM? Can your store answer a financing or trade question in the moment, or does each one still wait on a person? Does a single system hold the full picture of a customer, or is that picture spread across systems that never share it? The stores that answer easily are further along than they tend to think. The ones that struggle have simply found their work, and it’s worth beginning before the agent is the one asking.
What’s real today
It’s early, and early is when the advantage gets built. Our AI agents answer a shopper’s financing questions with real, individualized numbers inside the conversation itself, a soft credit pull and a payment estimate by term, with no handoff, on one of the harder parts of a car deal to automate.
The dealers who do well in this era won’t be the ones who bought the most point technology solutions. They’ll be the ones a machine can read, across everything their store does, the moment it comes looking. Selling has always decided who wins the deal. Legibility is what now decides who’s in it at all, in the showroom and the service drive alike. And that’s a question worth answering while the answer is still yours to shape.
Drive the Future breaks down the shifts reshaping automotive retail, one thesis at a time. Subscribe to get each issue.
_____________
Frequently asked questions
What is agentic commerce?
Agentic commerce is a model where an AI agent researches, compares, negotiates, and buys on a person’s behalf, rather than the person browsing and checking out manually.
What does agentic commerce mean for car dealers?
When an agent does the shopping, the store it can find, read, and deal with gets considered, and the one it can’t is skipped. Being legible to the agent becomes the new price of entry, before selling ability ever comes into play.
How does agentic commerce reach automotive?
Through the simplest transactions first. Routine, low-stakes purchases go agentic ahead of complex ones, so a financed vehicle purchase, with its trade, financing, and incentives, is the hard end of the shift, not the entry point.
How can a dealership get ready for agentic commerce?
Make sure your inventory, pricing, incentives, and customer history are connected, current, and machine-readable through one connected AI Operating System, so your store can answer real payment, trade, and financing questions inside the conversation rather than handing them off.