A quarter of new-car buyers now ask an AI before they ask you
Cox Automotive puts AI use at 25 per cent of new-car buyers. When the answer arrives without a click, being the source the AI quotes matters more than ranking.
Cox Automotive’s 2026 Car Buyer Journey Study found 19 per cent of all vehicle buyers, and 25 per cent of new-car buyers, now use AI tools during research. A separate study this year put the figure closer to 30 per cent, with ChatGPT accounting for around 68 per cent of AI-driven vehicle research. Of buyers who used AI, 88 per cent found it helpful, and 40 per cent of future buyers say they intend to use it next time.
The more interesting finding is what it does to conversion. Dealerships appearing in AI recommendations are reportedly seeing leads convert at around 4.4 times the usual rate — which makes sense. A buyer arriving from a recommendation has already been pre-qualified by something they trust.
Why this is not just another SEO channel
Traditional search gave you a shot at a click. The buyer saw ten results, formed their own view, and visited two or three. Being fourth was still worth something.
An AI answer works differently. The buyer asks “which mid-size electric SUVs under $60,000 are worth looking at in Australia” and gets a shortlist of three with reasoning attached. There is no page two. Either you are in the answer or you are absent from the consideration set entirely — and the buyer will never know you existed to be excluded.
Frequently they never click at all. Which means the traffic report understates what is happening: your visibility can be falling in analytics while your influence is intact, or your visibility can look stable while an AI is quietly recommending a competitor to a quarter of the market.
What actually gets a brand into the answer
These systems synthesise from sources they can read and are inclined to trust. In this category, in practice:
Specific, structured, answerable content. “Range, charging time and running cost for the [model], with figures” gets used. A brochure page of adjectives does not. The content that gets quoted looks like a reference document, not a campaign asset.
Real numbers on the page as text. Specifications locked inside an image, a PDF or a JavaScript-rendered tab are invisible. This is the single most common technical failure in automotive, and it is entirely fixable.
Comparison content, including unflattering comparisons. These systems are asked comparative questions constantly. A brand that publishes honest comparisons becomes a usable source; a brand that only publishes its own strengths gets treated as a brochure and skipped in favour of a review site.
Third-party corroboration. Reviews, owner forums, motoring press and marketplace listings all feed the answer. You cannot control these, but you can be accurate and consistent everywhere they draw from — inconsistent pricing and spec across your own site, your listings and your dealer sites actively costs you.
Structured data. Schema markup on vehicles, stock, pricing, FAQs and organisation details. It is not a magic switch, but it makes the content unambiguous, and unambiguous content is easier to quote.
What to do about it
- Test it. Ask the ten questions your buyers actually ask, across ChatGPT and Google’s AI results. Write down what comes back, who is named, and what is said about you. Most brands have never done this once. It takes an afternoon and it is frequently alarming.
- Fix the retrievability basics. Specifications as text, stock and pricing crawlable, no critical content behind interaction, schema in place.
- Publish the reference layer. Running costs, charging, towing, servicing intervals, warranty terms, fitment. The dull, factual content that most brands consider beneath them is exactly what gets cited.
- Audit for consistency. Same price, same spec, same availability across your site, your dealer sites and your marketplace listings.
- Measure it separately. Assisted and direct traffic from AI sources, tracked as its own line. If it sits inside “referral” you will never see the trend.
The honest caveat
This is a fast-moving area with a lot of confident vendors and not much settled measurement. Anyone selling you a guaranteed method for ranking in AI answers is guessing.
But the underlying work is not speculative. Making your factual content accurate, structured, readable and consistent is the same work that makes it useful to buyers, quotable by search engines, and easier for your own dealer network to rely on. That is worth doing whether or not the current AI numbers hold.