

Cazoo just rolled out a free AI tool for dealers on its Dealer Centre platform, built with customer-engagement vendor Kaisa. It records inbound calls, transcribes them, and spits out a summary flagging the stuff that actually matters: trade-in details, financing questions, deposit talk, requested callbacks. Cazoo's chief product officer framed it as a dealer-requested fix — "when every callback happens and every promise is kept, dealers sell more cars."
That's a fine sentence. It's also doing a lot of quiet work, because the tool doesn't make the callback happen. It just tells you the callback should have happened. Someone still has to read the summary, act on it, and log it. That's precisely the step where the industry's numbers say things fall apart.
What the German panel data actually shows
Autohaus.de's August 2026 dealer panel is worth more attention than most "AI adoption is up" stories, because it breaks down where the usage lives, not just that it exists. AI usage among German dealers jumped from 37% to 52% in a year — real growth. But only 4% use it deeply and deliberately across the business. Another 17% use it in a structured way in specific departments. The remaining chunk — nearly a third of respondents — admit they're chasing tools that show quick, visible wins rather than solving harder structural problems.
Marketing and social media lead adoption at 59%. Lead generation and customer communication follow at 52% — exactly the category Cazoo's tool sits in. Vehicle pricing and service reception trail behind. And tellingly, 57% of dealers are running general-purpose tools like ChatGPT or Copilot rather than automotive-specific software, meaning a lot of "AI in the dealership" right now is a sales manager pasting call notes into a chatbot, not an integrated system.
The financial signal is genuinely good, though: 41% of users say benefit already exceeds cost, and 29% call the return clearly positive. Only 9% say they're underwater. AI in dealerships isn't vaporware. It's paying for itself for people who use it with intent.
But here's the line that should worry every GM reading this: only 42% of dealers running AI tools have any written rules for how they're used — no approval process, no defined ownership, no governance. That's the same failure mode showing up in the lead-response numbers below, just one layer up the stack.
Where the leads actually die
Ventavid's 2026 roundup of lead-response statistics is the reality check. Dealers are demonstrably faster than they used to be — 61% now respond to an internet lead within 15 minutes, up from 55% in 2022, and the share giving a "perfect response" to a web lead has doubled in five years to 51%, per Pied Piper's mystery-shop data.
Progress, genuinely. But look at the other half of the ledger: 62.8% of sales leads got no salesperson follow-up within 24 hours after returning to the dealer's own website. And 15.2% of qualified leads — a number that's climbed for six straight years — never made it into the CRM at all. Not mishandled. Never logged. Gone.
That second stat is the one that should reframe how dealers think about tools like Cazoo's. A perfect AI summary of a great phone call is worthless if the lead it describes never gets a CRM record, or if the summary sits unread in a BDC inbox because nobody owns the follow-up. Call-intelligence tools solve the documentation problem. They don't solve the accountability problem — and the data says accountability is where dealers are actually bleeding deals.
The operator takeaway
None of this means skip the tool. Free AI call summaries that flag deposits, trade-ins, and callback promises are a genuinely useful layer, especially for stores without a disciplined BDC process today. But treat it as an input, not a fix. Before rolling it out, ask the harder question the German panel exposes: who owns turning that summary into a logged, tracked, time-stamped follow-up task — and what happens when they don't?
If the answer is "the salesperson will handle it," you already know how that ends, because 62.8% of leads say it doesn't. Buy the AI. Then go fix the handoff it's just going to keep exposing.

A new Pasch Group report on AI rollouts at 200+ dealerships finds the software matters less than whether anyone bothered to plan the launch, connect the systems, and write a policy. One store using the same AI as its group-mates redesigned its workflow and tripled its contact-rate gains — proof that the tool was never the bottleneck.

This week's evidence says the AI isn't the variable — your setup is. One dealership tripled its group's average contact-rate gain on identical software, while mystery shoppers found customers juggling four different chatbot "names" before a human ever picks up the phone.