The Agent-Era Car Shopper: What AI Buyers Mean for Your Dealership's Phones and BDC

Picture a quiet Tuesday on the desk. A salesperson opens what looks like an ordinary internet lead, except the wording is unusually precise: it names your exact stock number, references two competing quotes within $200 of each other, and asks for an out-the-door price plus the first open test-drive slot. A few minutes later, three nearby competitors get the same message. The shopper didn't write a word of it. Their AI agent did.

That scene is closer than most dealers think. Cox Automotive's analysts describe live demos of consumer AI agents that already research and build scenarios, secure pre-approval and set appointments, read an in-store four-square from a photo, and fire competing-offer requests to several stores at once. The useful question isn't whether agent-assisted buyers are coming. It's how much of your current process quietly assumes the person on the other end is moving at human speed.

This piece looks at what the data actually says about AI-assisted car shoppers in 2026 — how common they are, what they expect, and where the pressure lands first. The short version: it lands on your phones, your BDC, and your response time.

Myth: AI car shoppers are a 2030 problem you can deal with later. What the data shows: AI already factored into the journey for 19% of all car buyers — and 25% of new-vehicle buyers — in 2025, the first year Cox Automotive tracked it. Buyers who used AI also reported some of the highest satisfaction in the entire study. (Cox Automotive, 2025 Car Buyer Journey Study)

From "asking AI" to "AI doing the work"

There's an important distinction buried under the hype. A generative chatbot answers questions. An agent takes actions — it reasons, plans, and completes tasks across systems. Cox Automotive frames the difference plainly: chatbots talk with you; AI agents work for you. For a store, that's the gap between a shopper who shows up better informed and a shopper whose software has already pre-screened three rooftops and booked the test drive before anyone picked up the phone.

Car buying is unusually exposed to this shift. It's a high-consideration purchase with multiple quotes, financing, a trade, and a negotiation — exactly the repetitive, comparison-heavy work people are most eager to hand off. And the adoption curve, while early, is steep. Morgan Stanley Research estimates that "agentic shoppers" could account for $190 billion to $385 billion in U.S. e-commerce spending by 2030 — roughly 10% to 20% of the market — and notes that about 23% of Americans already made a purchase using AI in the prior month. Most of that is groceries and everyday goods today, not vehicles. But the habit is forming, and car shoppers aren't a different species of consumer.

The automotive-specific signals point the same way. Automotive News reported on a Massachusetts software developer who used an open-source AI agent to help negotiate his Hyundai Palisade purchase — a novelty today, a preview of behavior tomorrow. And in Cox Automotive's data, AI sites already registered as a measurable research channel, used by 12% of all buyers and 17% of new-vehicle buyers, sitting alongside the third-party and dealership sites buyers have relied on for years.

The takeaway isn't that your lot is about to fill with robots. It's that a growing share of your inbound — calls, forms, and chats — is being shaped, and increasingly initiated, by software that expects an immediate, accurate, and consistent answer.

Why the pressure hits your phones first

Agents are unforgiving in a specific way. They don't wait on hold, they don't leave a voicemail and hope for a call back, and they don't extend the benefit of the doubt. They move to the next store. That turns two long-standing dealership weak spots — slow response and inconsistent information — from quiet leaks into deciding factors.

The response-time math was punishing even before AI arrived. The widely cited Lead Response Management study led by Dr. James Oldroyd (2007) found that contacting a web lead within five minutes instead of thirty made it 21 times more likely to enter the sales process. Harvard Business Review's 2011 audit of 2,241 U.S. companies found the average first response took 42 hours, and 23% of firms never responded at all. If a human-paced lead already punished a slow reply, an agent that pings five dealers at once simply hands the deal to whoever answers first — and correctly. (We dig into the dealership version of this in speed to lead in automotive.)

Consistency is the second pressure point. Cox Automotive's research shows buyers increasingly expect to resume in-store exactly where they left off online — 85% expect that continuity — yet 97% of dealers admit customers end up repeating steps in the showroom. A human shopper might shrug that off. An agent that collects one quote by phone, a different figure from your website, and a third from a form does not; it flags the discrepancy and marks down your credibility.

The expectation gap agents will expose

Here's where it gets concrete for managers. Cox Automotive's 2025 data shows a persistent gap between what shoppers want to do online and what dealers actually let them do. Agents live in that gap — they thrive on steps that can be completed digitally and stall, or route around you, on the ones that can't.

Source: Cox Automotive 2025 Car Buyer Journey Study (2,300 U.S. buyers, fielded fall 2025).

Each of those gaps is a point of friction for an agent-assisted shopper. Dealers who close them aren't only lifting CX scores — they're making themselves legible to software that rewards clarity. And the payoff shows up with human buyers too: Cox found that shoppers who completed more steps online saved about 42 minutes in-store and were more likely to be highly satisfied (81% versus 71%).

What to do this quarter — not this decade

It's easy to over-rotate on a trend like this. The measured read is that agent-driven buying is still a small slice of demand, but a fast-growing one — and the moves that prepare you for it are the same ones that pay off with ordinary human shoppers today. A few the data supports:

Answer every inbound, every hour. The most agent-proof thing a store can do is make sure no call, form, or message goes unanswered — including nights and weekends, when an agent is perfectly content to work. That's the core case for AI coverage on the phones, and we lay out where it helps and where it doesn't in AI phone agents and the dealership scaling gap and whether an AI BDC can replace your human team.

Quote one number, everywhere. Decide where your single source of pricing truth lives, and make the phone, the website, and the showroom agree. Inconsistency is precisely what an agent is built to catch.

Treat a missed call as a lost deal, not a lost call. When a shopper — or their agent — can re-broadcast the same request in seconds, the cost of not picking up compounds quickly. We put numbers to that in the real cost of missed calls.

Compete on what an agent can't commoditize. If price-shopping can be automated in minutes, your edge shifts to trust, transparency, and certainty: a clear out-the-door figure, an accurate payment, and a test drive that's actually on the calendar. Cox Automotive's own conclusion is blunt — in the agent era, experience and relationship become the real differentiators.

None of this requires betting the store on a 2030 forecast. It requires assuming that a meaningful share of the people contacting you this quarter are already using AI to decide who's worth their time — because, by Cox's own count, roughly one in five of them is.

Dealers who treat speed and consistency as table stakes won't just be ready for agent-driven shoppers. They'll win more of the human ones along the way.

Carbuki builds AI voice agents that help U.S. dealerships answer every sales and service call — day or night — with fast, consistent, accurate information. If you're working through what the agent era means for your phones, see what we're building at carbuki.com.

Sources

• Cox Automotive, "Car Buyer Journey Study Finds Efficiency, Digital Tools and AI Drive Record Satisfaction" (Jan. 13, 2026): AI usage (19% of all buyers, 25% of new-vehicle buyers), AI-site channel share, satisfaction, and online-vs-actual gaps. https://www.coxautoinc.com/insights/cox-automotive-car-buyer-journey-study-finds-efficiency-digital-tools-and-ai-drive-record-satisfaction/
• Cox Automotive Deal Central, "The New Dealer's Edge: From Chatbots to AI Agents" (2026): agent-vs-chatbot framing, the four agent moments, the 85% / 97% continuity gap, and 42 minutes saved. https://www.coxautoinc.com/deal-central/resources/the-new-dealers-edge-from-chatbots-to-ai-agents-and-a-truly-seamless-click-to-keys-journey/
• Morgan Stanley Research, "Here Come the Shopping Bots" (Dec. 8, 2025): agentic shoppers projected at $190B to $385B of U.S. e-commerce by 2030; about 23% of Americans bought via AI in the prior month. https://www.morganstanley.com/insights/articles/agentic-commerce-market-impact-outlook
• Automotive News, "Dealerships face new challenge: Buyers with custom AI agents" (2026). https://www.autonews.com/retail/an-ai-produces-new-kind-of-car-customer-0319/
• J. Oldroyd, Lead Response Management Study (2007); J. Oldroyd, K. McElheran and D. Elkington, "The Short Life of Online Sales Leads," Harvard Business Review (2011): the five-minute response advantage; 42-hour average first response; 23% never respond. https://hbr.org/2011/03/the-short-life-of-online-sales-leads