Your Next Customer May Never Visit Your Website

A customer wants a new SUV.

Not long ago, we had a pretty good idea how that shopping process would start.

Google. Autotrader. Cars.com. The OEM website. A dealer website.

Eventually the customer lands on a VDP, checks inventory, looks at pricing, reads reviews and maybe submits a lead.

We built an entire automotive retail ecosystem around capturing that shopper.

SEO. Paid search. Third party leads. Retargeting. Website conversion. CRM attribution.

Almost everything depends on getting the customer into an environment we can see.

Now imagine the same customer opens ChatGPT or Gemini and types:

Find me a Lexus RX 350 AWD near Detroit. I want Premium or Premium Plus, black exterior, under $60,000. Compare the dealers based on inventory, price, reviews and reputation and tell me where I should go.

The customer may still end up on a dealership website.

But they may only visit one.

The AI already helped eliminate the other four.

That is the part I think dealers should start paying attention to.

We have spent the last twenty years fighting over who gets the click.

The next fight may be over who gets recommended before the click ever happens.

Cox Automotive released its first AI in Auto Retail Tracker this month, and one number immediately stood out to me.

63% of shoppers say they will definitely or probably use AI during their next vehicle purchase.

At the same time, only 29% of dealerships have started adjusting to AI powered search.

That is a pretty large gap.

It gets more interesting when you look at dealer adoption overall.

82% of dealers already say they are using AI.

So this is not really a story about dealerships refusing to use artificial intelligence.

Most of us already are.

The issue is what we are using it for.

According to Cox, the most common dealer uses today are automating tasks, coordinating customer follow up and generating content.

Basically, we are using AI to become more efficient once the customer is already somewhere inside our ecosystem.

That makes sense.

But shoppers are beginning to use AI further upstream, before they ever become our lead.

And that changes the transaction in a much bigger way.

Look around the industry right now and everybody has an AI product.

AI receptionist.

AI BDC.

AI CRM assistant.

AI service scheduler.

AI equity mining.

AI pricing.

AI lead response.

AI call analysis.

AI training.

Some of these products are genuinely useful.

But most of the conversation inside dealerships has been about using AI to improve our operation.

Meanwhile, the customer is getting their own AI.

That customer can ask:

What vehicle should I buy?

What trim is worth the money?

What should this car actually sell for?

Which dealership has the best price?

Which dealership has the best reviews?

What should my trade be worth?

What incentives are available?

Should I lease or finance?

What should I ask the salesperson?

Is this quote a good deal?

That customer can walk into the dealership having already had an hour long conversation about the transaction without speaking to anyone at the dealership.

I do not think that removes the salesperson.

I think it changes the customer the salesperson meets.

They may arrive later in the decision process.

They may already understand the trim levels.

They may already know competing inventory.

They may already have an idea of market pricing.

And increasingly, an AI may have helped decide which dealership made the list.

This is where things get interesting from inside the store.

Imagine five dealerships within driving distance all have the vehicle.

All five have websites.

All five have competitive pricing.

All five spend money on Google.

All five push inventory across the major marketplaces.

The customer asks an AI where they should buy.

It recommends three.

The customer visits those three dealership websites, contacts two and eventually buys from one.

What happened to the other two stores?

Nothing.

There was no lead to lose.

No phone call to mishandle.

No appointment that did not show.

No VDP visitor that failed to convert.

No CRM record.

Those dealerships simply never made the shortlist.

From their perspective, that shopper never existed.

That is a very different type of lost customer than dealers are used to measuring.

For years we have obsessed over what happens after someone finds us.

Did we answer quickly?

Did we respond correctly?

Did we get the appointment?

Did the salesperson follow up?

Did the manager get involved?

All of that still matters.

But AI introduces another question.

Did we even make it into the conversation?

Pied Piper added a ChatGPT test to its 2026 Internet Lead Effectiveness study.

The setup was simple.

They prompted ChatGPT as a local shopper asking about a specific vehicle that was known to be in stock at a nearby dealership.

The target dealership appeared first 77% of the time.

That sounds pretty good.

But 14% of the time, the dealership appeared without being ranked first.

And 9% of the time, the dealership was not mentioned at all.

Remember the setup.

The dealership was nearby.

The dealership had the requested vehicle.

And the AI still did not include it.

ChatGPT correctly confirmed that the vehicle was in inventory 70% of the time.

Only 30% of responses included a direct link to the exact vehicle.

That should get a dealer's attention.

We are used to thinking that if the car is online, the customer can find it.

That assumption is starting to weaken.

You can have the right car, at the right store, in the right market and still lose visibility before the customer ever sees the inventory.

There is going to be a lot of bad advice around this.

We will hear that SEO is dead.

That dealership websites are dead.

That Google is dead.

That dealers need some new AI optimization product immediately.

I would be careful with all of it.

Google's own guidance around AI search says the fundamentals still matter.

Pages need to be crawlable.

Important information needs to exist in readable text.

Structured data needs to match what is actually on the page.

Business Profile information needs to be accurate.

Content still needs to be useful.

There is no magic AI tag that suddenly makes a search engine recommend your dealership.

So I do not think the website becomes less important.

I think its job expands.

We built dealership websites mainly for humans.

Now they also need to make sense to machines acting on behalf of humans.

That is a subtle difference.

But a big one.

Think about how dealership websites are normally designed.

Huge banners.

Popups.

Payment widgets.

Trade tools.

Chat boxes.

Coupon offers.

Lead forms.

Sticky buttons.

Another popup.

Somewhere underneath all of that is the actual information the customer came for.

Now imagine an AI trying to understand the dealership.

What vehicles are actually available?

What is the real price?

What packages does the vehicle have?

Is there a market adjustment?

Can the vehicle be purchased today?

What are the store hours?

Does the dealership sell out of state?

Does it deliver?

What fees are added?

What incentives apply?

What does the service department offer?

What do customers consistently say about the store?

That information needs to be clear.

Not only visually clear to the customer.

Structurally clear enough that search engines and AI systems can understand it.

That makes some of the most boring work in the dealership much more important.

Accurate inventory.

Correct prices.

Clean VDP information.

Updated hours.

Useful vehicle descriptions.

Consistent dealership information.

Accurate Google Business Profiles.

Real customer reviews.

Clear service information.

Clear policies.

None of that sounds like artificial intelligence.

But that may be exactly what determines how well artificial intelligence understands your dealership.

I think reviews are another area dealers should watch closely.

Historically, we have treated reviews mainly as a rating.

4.8 stars.

4.6 stars.

4.3 stars.

Get more five star reviews and push the bad ones further down.

AI can change how customers use that information.

Instead of reading 200 reviews, the customer can ask:

Which dealership is known for being the most transparent?

Or:

Which Lexus dealer near me has the best reputation for service after the sale?

Or:

Do customers complain about hidden fees at this dealership?

An AI can potentially digest a large amount of public information and summarize the patterns.

That changes the importance of what customers actually say.

Not just the star rating.

Transparent pricing.

Easy transaction.

No pressure.

Good communication.

Hidden fees.

Long waits.

Poor follow up.

Great service department.

Those themes become data.

And if customers increasingly ask AI to evaluate dealerships, reputation stops being only something customers see.

It becomes something machines interpret.

Dealers love metrics.

Website sessions.

VDP views.

SRP views.

Lead conversion.

Cost per lead.

Cost per sale.

Engagement.

We have dashboards for dashboards.

But consider what happens if AI handles more of the research.

The customer asks ten questions.

AI narrows ten models down to three.

Then identifies four available vehicles.

Then compares three dealerships.

Then the customer finally clicks one VDP.

Our analytics see one VDP view.

But that customer may have already completed half the shopping process before they arrived.

That could make raw traffic less meaningful over time.

Someone who searches Lexus RX is very different from someone who arrives after asking an AI twenty questions about an RX 350 Premium AWD and then clicks the exact vehicle they want.

Same website session.

Completely different customer.

Eventually, we may need to care less about how many people came through the front door and more about whether we were part of the consideration process that got them there.

I would not overcomplicate this.

You do not need another dashboard tomorrow.

You do not need to hire an AI search consultant.

You do not need to rebuild your marketing strategy.

Start simple.

Open ChatGPT.

Open Gemini.

Use Google AI Mode.

Then shop your own market like a customer.

Do not search the dealership name.

Ask:

Where should I buy a new Silverado near me?

Find me a used RX 350 under $45,000 near Detroit.

Which Toyota dealership near me has the best reputation?

Find me a RAV4 Hybrid that is available now.

Which store would you recommend and why?

Then see what happens.

Does your dealership appear?

Does your inventory appear?

Is the information accurate?

Does the AI understand what makes your store different?

What does it say about your reputation?

Which competitors keep appearing ahead of you?

I would be doing this monthly if I ran a dealership.

Not because the answers are perfect.

They are not.

AI search is still evolving quickly.

But I would want to know what customers might be seeing.

We mystery shop our BDC.

We mystery shop our salespeople.

We mystery shop competitors.

Maybe it is time we start mystery shopping the machines too.

There is another implication here.

Dealers have spent years publishing generic SEO content.

Five Reasons To Buy An SUV.

Benefits Of Leasing Your Next Vehicle.

Why The Toyota Camry Is A Great Sedan.

AI does not need us for that anymore.

Neither do customers.

If an AI can create essentially the same article in ten seconds, there is not much reason for a search engine to believe your dealership's version deserves special attention.

Dealers actually have something much better.

Real information.

What questions are customers asking every day?

Which trims are hardest to get?

What options are worth paying for?

What do buyers misunderstand about leasing?

Which used cars disappear immediately?

What problems are customers having with the buying process?

What does winter ownership look like in Michigan?

What should a first time Lexus buyer understand before choosing between an NX and RX?

That is information coming directly from the rooftop.

AI does not sit at the desk.

It does not appraise trades.

It does not talk to twenty customers every Saturday.

It does not see why the same deal keeps blowing up at $750 a month.

Dealers do.

That experience is becoming more valuable, not less.

Right now, we are mostly talking about AI helping customers research.

I do not think it ends there.

Today the customer asks:

Which dealership should I visit?

Eventually they may ask:

Contact the three best dealerships and get me pricing.

Then:

Compare the offers.

Then:

Ask them what they will give me for my trade.

Then:

Get the payment under $700.

Then:

Book the appointment with whoever has the best deal.

We are moving toward a world where the first interaction between a customer and a dealership may not involve two humans.

It may be the customer's software talking to the dealership's software.

That is when things get really interesting.

Because dealerships have spent decades training people to handle customers.

Very few have started thinking about how the store presents itself to machines acting for customers.

That may eventually affect inventory, pricing, lead attribution, response time, website architecture, CRM strategy, reviews, advertising and even desking.

We are not there yet.

But the direction is getting pretty obvious.

I do not think AI kills the dealership website.

I think it changes what the website is for.

The website is still where customers may shop inventory, verify information, submit leads and eventually complete transactions.

But increasingly, it may also become the information source another system uses to decide whether your dealership deserves to be recommended.

That means the battle begins earlier.

Before the VDP.

Before the lead.

Before the phone call.

Before the appointment.

Potentially before the customer even knows your dealership exists.

For years we asked:

Can the customer find us?

I think dealers need to start asking another question:

Does the machine understand us well enough to recommend us?

Because your next customer may still buy a car from you.

They just may not discover you the way the last one did.

What are you seeing inside your store? Are customers starting to mention ChatGPT, Gemini or other AI tools during the buying process? Have you searched your own inventory through AI yet? Follow Rooftop Insider on LinkedIn and join the conversation.

I'm Kamil Grzych. I work inside automotive retail, with experience across dealership sales, BDC operations, digital retail, customer experience and technology implementation across Toyota, CDJR and Lexus dealerships.

I started Rooftop Insider because I wanted something written from the dealership side of the desk.

Not another publication explaining automotive retail to people who already live it every day.

Rooftop Insider is an independent automotive retail publication focused on the forces changing the dealership business.

AI. Technology. Inventory. Pricing. Digital retail. Customer behavior. Operations. And the business models quietly changing how cars are bought and sold.

The goal is simple.

Take what is happening across automotive retail, separate the signal from the noise and explain what it actually means for the people running and working inside dealerships.

Rooftop Insider articles are based on independent analysis and real world dealership experience. No company referenced in this article paid for or approved its inclusion.

Sources: Cox Automotive AI in Auto Retail Tracker, August 2026; Pied Piper 2026 Internet Lead Effectiveness Study; Google Search Central guidance on AI search and generative search.