AutomotiveCase study3 min read
Automotive: the test drive booked while the showroom was shut
Buyers shortlist cars at midnight. The dealership that answers the odometer question is the one they turn up to.
The business in this scenario: An independent used-car dealership carrying around eighty vehicles
In short
Car buyers shortlist at midnight and ask about odometer readings, service history and whether a vehicle is still there. The dealership put an assistant on its site that answers from live stock records, raises trade-in and finance while the buyer is still interested, books the test drive, and attributes the enquiry to the campaign that produced it.
Nobody buys a car during business hours
They shortlist at midnight. A buyer opens six tabs across four dealerships, and works through the same three questions on each: what is the odometer reading, what is the service history, and is it actually still there.
That last one is the killer. Stock turns over daily and websites do not, so a serious buyer has learned to assume a listing might be stale. Asking is the only way to find out — and the dealership that answers is the one that gets the Saturday morning.
For a dealership this size the pattern is visible in the analytics and invisible everywhere else. Traffic peaks between ten at night and one in the morning. Enquiries arrive through a contact form. Replies go out at ten the following morning, by which time a good proportion of those buyers have already committed their weekend to somebody else.
Three problems, and one of them is expensive
The timing problem is the obvious one, and it is real: a midnight stock question answered at ten the next morning is a question answered too late to be worth much.
The second is sequencing. Trade-in and finance are the two facts that decide whether a deal is a deal, and both surface on the forecourt. A salesperson spends an hour with somebody who was never going to be financeable, or discovers a trade-in at the moment of handshake with no valuation prepared. The information exists; it is simply arriving at the worst possible point in the process.
The third is attribution, and it is the one that quietly costs the most. Marketplace listings, paid search and paid social all bill separately and all claim credit enthusiastically. None of them can tell you which one put a particular buyer in a particular seat, which means the monthly spend decision gets made on instinct dressed up as data.
Trade-in and finance are the two facts that decide the deal, and both surface on the forecourt — an hour too late to be useful.
Setup, in full
The assistant crawls the dealership site and takes the vehicle catalogue — make, model, year, odometer, price and status per unit. That catalogue is the thing that matters here: because the assistant answers from the records rather than from page copy, moving a car to Reserved or Sold is what changes what it tells the next buyer.
Then the paperwork buyers always ask for: finance terms, warranty documents, the trade-in process. Some of it is scanned rather than digital, which the ingestion handles.
Then one line of code, verified live by the install assistant. Total elapsed time is an afternoon, and at no point is a developer involved.
What a person does
- Let it read your stockIt crawls your site and takes your vehicle catalogue — make, model, year, odometer, price, status.
- Add the paperwork buyers ask forFinance terms, warranty documents, the trade-in process. PDFs and scans both.
- Paste one line of codeOr point the install assistant at your platform and let it verify the widget itself.
What the system does from then on, unattended
- Answers make, model, year, odometer, price and whether a vehicle is In stock, Reserved or Sold.
- Raises trade-in and finance early, in conversation, instead of on the forecourt.
- Scores and tiers each buyer with an AI summary and the full transcript attached.
- Books the test drive and drops the lead into the pipeline, assigned, with a follow-up date.
- Attributes the enquiry to the campaign that produced it — direct and influenced.
Twenty past midnight, on a Sunday
Here is the shape of the conversation that replaces the contact form. Note where the trade-in question appears — third exchange, unprompted by any human.
- Visitor: Do you still have the 2021 Corolla with under 60,000km?
- Assistant: Yes — 2021 Corolla, 54,200km, $22,400, in stock. Would you be trading anything in?
- Visitor: Maybe a 2015 Civic. Could I come and see it Saturday?
- Assistant: Saturday works. Morning or afternoon? I will note the Civic so the team has a valuation ready when you arrive.
- Specific vehicle identified
- Trade-in likely
- Test drive Saturday
Assigned to a salesperson eight hours before the showroom opened.
What that buys the dealership
The stock answer is live. The odometer figure, the price and the availability came out of the vehicle record, so they are as current as the dealership's own inventory rather than as current as whoever last edited the website.
The trade-in surfaced eight hours before the buyer arrived, which is the whole point. A valuation can be prepared. The salesperson walks into the conversation already knowing the shape of the deal instead of discovering it at the handshake.
The test drive was booked inside the conversation rather than through three rounds of phone tag, and the lead landed in the pipeline scored, tiered, assigned and dated, with the full transcript attached and the originating campaign tagged. When the month closes, the spend decision has a number behind it.
The shape of the change
Stated as behaviour rather than as a percentage, because behaviour is checkable.
- A midnight stock question waits for opening hours.It is answered from live stock data, at midnight.
- Trade-ins are discovered on the forecourt.The valuation is being prepared before the buyer arrives.
- Test drives are booked by phone tag.Booked in the conversation, staged in the pipeline.
- Every channel claims the sale.Cost per lead and ROAS, per campaign, against won revenue.
What to watch once it is running
The dashboard widgets are generated from the vehicle records, so they speak in stock terms from day one.
- Average price, year spread and stock mix, generated from your own vehicle records.
- Leads by tier, and how many convert from Hot to a completed test drive.
- Cost per lead by marketplace, paid search and social — with influenced attribution, not just last click.
- Follow-up compliance across the team, from the audit trail.
- Line of code to install
- 1Line of code to install
- Guided setup, start to live
- ~5 minGuided setup, start to live
- Languages, right-to-left included
- 30+Languages, right-to-left included
- Answering, including 2am
- 24/7Answering, including 2am
Product facts, checkable in the app. Not outcome claims.
Questions this industry always asks
- What happens when a car sells?
- Move the record to Sold in ConGreeto. The assistant answers from those records, so its answers move with your stock rather than with your website copy.
- Can it quote finance?
- It answers from the finance documents and Q&A you give it. Anything you have not taught it, it collects as a question and passes to the assigned salesperson with the transcript.
Keep reading
- Clinics & healthClinics and health: the enquiry that used to end in voicemailPatients want price, duration and who performs the treatment before they will book. Reception is on another call.
- Professional servicesProfessional services: qualifying before the discovery callHalf of a firm's discovery calls end in “not a fit”. The assistant establishes that before the calendar invite goes out.
- E-commerceE-commerce: the pre-purchase question that decides the cartShoppers abandon over one unanswered question about sizing, stock or shipping. Nobody was awake to answer it.
Your visitors are already asking these questions.
One line of code, about five minutes of setup, and something is there to answer them.