The Invisible Journey · Article 1 · While You Debate AI, Your Customer Is Already Using It
Automotive executives are having an important conversation about AI.
Which models should we use? Which agents should we deploy? Which workflows should we automate? Should we build internally, buy from a vendor, let the agency handle it, connect ChatGPT, connect Claude, wait for Gemini, or commission a committee to produce a color-coded spreadsheet comparing all of them?
Reasonable questions.
Your customer has another one:
Which AI can help me figure out this car purchase before I have to talk to anybody?
That is not quite the same thing as saying customers want to eliminate dealerships. The newest research suggests something more nuanced—and strategically more important.
Customers are increasingly using AI to remove uncertainty before they decide which dealership interaction is worth having.
They can use an assistant to compare vehicles, understand reliability, work through tradeoffs, estimate price ranges, formulate questions, investigate ownership costs, evaluate a trade, organize financing considerations, and build enough confidence to enter the next step of the journey far better prepared than the customer who walked onto a lot twenty years ago asking, “So what’s the difference between these two?”
By the time the dealership sees the lead, the website session, the phone call, or the person walking through the door, a meaningful portion of the shopping process may already have happened somewhere the dealership could not see.
While the dealer is deciding which AI to use, the customer is increasingly using AI to decide which dealer deserves the next interaction.
That changes the assignment.
The first AI challenge for automotive retailers may not be getting employees to use artificial intelligence faster.
It may be making sure the customer’s artificial intelligence has enough trustworthy evidence to understand why your business—and your people—deserve consideration in the first place.
The traditional automotive funnel has always contained an invisible portion. Customers talk to friends, notice vehicles in parking lots, read reviews, argue with spouses, remember a bad service experience, and decide what they can afford long before they submit anything resembling a lead.
What is changing is the sophistication of the infrastructure available during that invisible stage.
The 2026 Urban Science Harris Poll Study found that 44% of U.S. auto buyers say they now spend more time researching before contacting a dealership than they did a year ago. Thirty-one percent say they are putting greater emphasis on price and value over brand loyalty, while 25% report being more likely to buy a vehicle fully online.
Those are meaningful changes because they describe a customer taking more control of the journey before dealership contact begins.
But the same study gives traditional retail an important reason not to panic. Nine in ten U.S. buyers would still consider purchasing from a traditional dealership in person, and 66% would consider completing the transaction fully online through a traditional dealership’s website.
The dealership is not disappearing from the journey.
Its role is changing inside the journey.
Urban Science’s channel data illustrates this especially well. Dealer websites are important throughout research, but their importance grows as the shopper approaches purchase. Buyers report relying on dealer sites more heavily during comparison and narrowing, and dealer websites become the dominant automotive website category as customers prepare to buy.
That suggests the physical dealership and dealer-controlled digital properties increasingly operate as validation and action surfaces after substantial upstream research has already occurred.
The dealership may still be where the transaction becomes real. It is increasingly not where the customer’s understanding of the transaction begins.
This is the first strategic adjustment leadership teams need to make.
We have spent decades building technology around the moment the customer raises a hand. Leads, attribution, CRM workflows, response-time reporting, nurture sequences, call tracking, digital retail completions, appointment metrics—all incredibly useful once the customer becomes visible.
AI is dramatically increasing the amount of useful work the customer can complete before visibility begins.
The pre-lead journey is becoming an intelligent journey.
Cox Automotive’s newest research makes the pace of the shift difficult to dismiss.
Its August 2026 AI in Auto Retail Tracker found that 63% of in-market vehicle shoppers say they definitely or probably will use AI during their next vehicle purchase.
Twenty-four percent already say AI helps them feel better prepared when working with a dealership.
That last statistic deserves more attention than it will probably receive.
The customer is not simply using AI to retrieve an answer. They are using it to improve their own position in the transaction.
They can ask questions privately before exposing uncertainty to a salesperson. They can interrogate unfamiliar financial concepts without worrying that somebody is steering them toward a payment. They can compare two vehicles using the priorities that actually matter to their family rather than whatever filters happen to exist on a website. They can arrive with a list of questions generated from hours of research compressed into one conversation.
OpenAI describes the same behavior across retail more broadly. Its shopping research experience is explicitly designed to reduce the work involved in moving through dozens of sites, comparing products, understanding tradeoffs, refining preferences, and deciding what fits a shopper’s needs and budget.
In March 2026, OpenAI expanded that direction with richer product discovery in ChatGPT, describing a shopping experience where consumers can explore, compare, refine choices conversationally, and move from a vague need toward a narrower decision without reproducing the traditional sequence of browser tabs.
The automotive implication is straightforward.
The customer does not need to begin with:
Honda Pilot versus Toyota Grand Highlander.
They can begin with the actual problem:
We have three kids, one still in a car seat, grandparents ride with us twice a month, we drive to North Carolina twice a year, my wife hates driving anything that feels enormous, and I’d rather not spend $65,000. What should we actually look at?
That is a much richer expression of intent than most automotive websites have historically been designed to receive.
And the AI can help the customer translate that intent into a shortlist before a dealership knows the shopper exists.
This is where I would challenge the most alarmist interpretation of the trend.
The evidence does not support a simple thesis that shoppers are adopting AI because they want dealerships removed from the process.
Cox explicitly found that avoiding dealerships ranks among the less common motivations for using AI. Instead, shoppers describe AI as a way to research vehicles, develop questions, understand choices, and become better prepared for the conversations they still expect to have with dealership staff.
Cars.com’s consumer research points in the same direction. Its study found that 73% of AI users see conversational AI as a time saver during vehicle research, with common use cases including identifying models that fit a shopper’s needs, comparing vehicles, understanding pricing, and researching issues such as reliability.
Then something important happens.
After using AI, 41% of respondents said their likely next destination was a cited dealer or manufacturer website.
The AI does not necessarily eliminate the dealer.
It can decide which dealer gets invited into the next step.
The customer is not using AI simply to avoid people. They are using AI to avoid wasting time with people who have not yet earned the interaction.
That changes the value of the human moment.
The customer who eventually engages may arrive with a more developed understanding of the market, clearer expectations, more specific objections, and a stronger sense of what an acceptable experience should look like.
They may also arrive carrying assumptions created by information the dealership did not provide.
Some of those assumptions will be accurate.
Some will not.
The quality of the dealership’s digital evidence therefore matters before the customer ever reaches a salesperson.
This is the question I would put in front of every dealer principal, OEM executive, and agency leader right now.
Not simply: Are we showing up in ChatGPT?
Ask instead:
What version of our business can an intelligent system actually construct?
Can it understand your identity?
Does it know which brands and communities you serve? Can it find accurate inventory? Does it understand your service capability? Are your people identifiable? Is there evidence of what your technicians, advisors, salespeople, and managers actually know? Are your reviews consistent with the experience you claim to provide? Are policies, offers, vehicle facts, and operating information coherent across surfaces?
Most importantly, when a customer asks a nuanced question, does the digital evidence surrounding the dealership reveal anything meaningfully different from the fifteen stores around it?
This is where automotive’s obsession with generated content becomes potentially dangerous.
If every dealership can create technically competent generic information instantly, generic information stops distinguishing anyone.
The scarce input becomes the part the generic model cannot independently invent: firsthand expertise, specific local knowledge, real customer experience, identifiable people, trustworthy inventory, institutional history, credible reputation, and useful answers grounded in the reality of the business.
The challenge is no longer making AI sound more human. It is making the humanity already inside the dealership visible enough for AI to understand.
A master technician’s explanation of a recurring ownership issue is evidence.
A salesperson who has helped hundreds of families make the same difficult comparison possesses evidence.
A service advisor who can clearly explain why a recommended repair matters possesses evidence.
A thoughtful video showing the actual tradeoffs between two trims is evidence.
A customer’s description of how they were treated during a difficult service event is evidence.
The dealership needs systems capable of capturing those signals, structuring them, validating the supporting facts, and allowing the resulting knowledge to travel.
That is no longer simply content marketing.
It is trust infrastructure.
There is another tension inside Cox’s new data that deserves executive attention.
Eighty-two percent of dealers report that they already use AI.
That sounds encouraging until we look at what they are primarily using it for.
The most common applications include task automation, customer follow-up, and content generation.
Those can all create value. But the customer is increasingly applying AI to the other side of the marketplace: research, comparison, preparation, evaluation, and deciding how to proceed.
Only 29% of dealers in Cox’s study had begun adapting specifically to AI-powered search, while another 32% recognized the need but had not yet started.
There is the strategic asymmetry.
The dealer is asking AI to help operate the dealership.
The customer is asking AI to help evaluate the dealership.
Both matter.
But they create very different competitive advantages.
A faster internal workflow is useful. A lower content-production cost is useful. An automated follow-up process may be useful when designed properly.
None of them guarantees that the dealership becomes part of the customer’s consideration set upstream.
That requires a different type of readiness.
The business has to project enough reliable context into the environments where intelligent systems are helping customers make decisions.
This is where the traditional idea of “digital presence” needs to mature.
For years, automotive strategy assumed that the objective was to attract customers toward a destination the dealership controlled.
Rank the webpage.
Generate the click.
Capture the lead.
Retarget the shopper.
Get the appointment.
Those mechanics still matter.
But the emerging journey is more distributed than that model allows.
The customer may be talking to ChatGPT while watching YouTube, comparing a Cars.com listing, checking Google reviews, scanning a dealership website, revisiting an employee they saw on social, asking Gemini another question, and eventually deciding that one dealership appears more knowledgeable, trustworthy, or relevant than the others.
The dealer cannot control the entire environment.
It can influence the quality of the evidence available inside it.
This is the new challenge for sophisticated dealers, OEMs, and agencies:
Project the truth, expertise, identity, and humanity of the business everywhere they need to live so the customer can build trust however they choose to shop.
That requires more than syndicating inventory.
It requires a coherent knowledge environment.
OEM vehicle truth has to coexist with local dealer truth. Dealership identity has to coexist with employee expertise. Content intelligence has to coexist with inventory and reputation. Agencies need the ability to contribute without reconstructing the brand in every prompt. Customers need increasingly consistent answers across the interfaces they choose.
And none of this works particularly well if each intelligent interface has to independently scrape the open web and guess.
This is where Hrizn’s v6 interconnectivity work becomes strategically more important than the phrase “AI integration” implies.
Hrizn MCP extends the dealership’s Content Operating System into AI environments including Claude, Cursor, ChatGPT, Gemini, and other MCP-compatible clients.
The point is not that Hrizn picked the “right” assistant.
The point is that the dealership should not have to.
Authorized assistants can work from the same governed store—the same Dealer DNA, brand voice, staff information, IdeaCloud research, live inventory, content intelligence, market context, OEM compliance logic, social workflows, and operating rules already maintained inside Hrizn.
Instead of every new AI rebuilding its own approximation of the dealership from whatever it happens to find online, the intelligence layer can connect to a more coherent source of business context.
That architecture matters for dealers because their preferred tools will change.
It matters for agencies because the agency desk and the rooftop should not operate from competing versions of reality.
It matters for OEMs because national vehicle and brand intelligence should be capable of flowing into local execution without eliminating the expertise, identity, and customer context that make an individual dealership relevant.
And it matters because the broader technology ecosystem is moving toward the same model. The emerging future is not one AI owning every workflow. It is increasingly a network of interoperable intelligence, governed context, tools, protocols, and specialized systems capable of collaborating around authoritative knowledge.
The moat is not owning the customer’s AI. The moat is becoming the most trustworthy business for any authorized AI to work with.
This is the collaborative superhighway we think automotive needs.
Not a new wall around Hrizn.
A governed way for dealership intelligence to travel.
Because the customer has already started traveling without us.
In the next article, we follow that journey farther upstream and ask the uncomfortable question:
What happens when the first meaningful conversation about your dealership takes place before you know there is a customer?
Next: The First Conversation Is Happening Without You. →
Return to the series hub: The Invisible Journey: Your Customer Is Already Shopping Without You. Make Sure Your Expertise Isn’t.
2026 Urban Science Harris Poll Study →
A detailed look at the increasingly fragmented, research-heavy automotive buyer journey and the changing role of dealers, OEMs, and digital channels.
Cox Automotive AI in Auto Retail Tracker →
Current data on the widening gap between shopper adoption of AI and dealership adaptation to AI-powered discovery.
Where Car Buyers Actually Search in 2026 →
See how vehicle discovery is spreading across search, AI, social, video, marketplaces, dealer sites, and physical retail.
The Human Signal Surge →
Why identifiable people and firsthand expertise become more valuable as generic information becomes abundant.
Get Your Dealership Cited by AI →
Practical guidance for building the evidence intelligent systems need to discover, understand, and reference a dealership.
As work spreads across assistants and agents, the dealership should not have to rebuild its identity, inventory context, research, voice, compliance rules, and content intelligence inside every new tool.
Hrizn MCP lets authorized AI clients work from the same live Content Operating System already shared by the rooftop and its partners—so the intelligence can move without the business losing control of the context underneath it.
One governed store. Many intelligent interfaces. Let the expertise travel.
Hrizn v6 helps dealership marketing teams move from fragmented activity toward a connected operating advantage—bringing intelligence, creation, human participation, distribution, proof and improvement into a more coherent system.
The next phase is not about forcing customers, employees, dealers, agencies, or OEMs into one AI environment. It is about giving trustworthy dealership intelligence the ability to meet them inside the environments they choose.
We Rise Together.
