Case Study May 11, 2026 6 min read

Six Days from Email to Sold

Saturday morning, May 9th. Brianna walked onto the lot a full week before her booked appointment, signed the paperwork on a 2018 Chevy Silverado, and drove off with the keys. Six days earlier she'd never spoken to a person at House of Carz. The only one she'd corresponded with was Jordan — an AI BDC I built from scratch. Here's the story, and the math.

Brianna standing next to her newly-purchased 2018 Chevrolet Silverado in front of the House of Carz dealership in Rochester, Indiana, holding the paperwork from her purchase on May 9, 2026.

Brianna with her 2018 Chevy Silverado, taken at House of Carz the morning she picked it up — May 9, 2026.

Three weeks ago, I flipped Jordan live at House of Carz.

If you read the first post, you know the why: a regional independent dealer cannot afford a full-time BDC team, but a regional independent dealer absolutely cannot afford to let a Sunday night lead sit until Tuesday morning either. The math doesn't work either way. So I built the third option.

For three weeks I've been watching Jordan answer customer emails, take inbound calls, schedule appointments, and follow up on dropped threads. Watching every word, mostly. Tuning the prompts. Fixing the bugs we'd built our way into. It was working — appointments were being booked, customers were replying, the inbox was full of conversations that read like a real BDC rep on a good day.

But I hadn't seen a sale. Not one I could honestly trace back to Jordan as the first-touch.

Until Saturday.

By the numbers

6
Days from first email to sold
$22,995
Sale price
$4,800
Total gross profit
0
Phone calls before she walked in

The truck: a clean 2018 Chevrolet Silverado 1500 we'd had on the lot for a few weeks. Stock #2044. Twenty-two nine ninety-five asking. The customer — let's call her Brianna, because that's her name and she said it was okay to share — found us through one of our online listings and submitted a lead form on a Sunday afternoon.

The arc, day by day

Sunday, May 3 — Day 0
Brianna's lead lands. Within seconds, Jordan sends his first email — a real one, in our small-town-Indiana voice, with a clean spec card for the Silverado and a call-to-book. She replies the same evening, asking if it's still available. Jordan confirms. She asks for the soonest test drive. Jordan offers options. She picks Saturday the 16th at four-fifteen. Jordan books her in.
Sunday, May 3 — same evening
The booking exposes two bugs in our scheduler — a Postgres array-literal quirk in our Knex queries, and an AI re-call idempotency issue that would have double-booked the slot. I find both that night. Both shipped within the hour. Brianna's appointment ends up the canary that flushes them out, but her record stays clean.
Monday, May 4
A day later, Jordan tries to send a generic re-engagement follow-up. It comes out odd — out-of-context for someone who'd already booked. I catch it. Spend the morning rewriting the re-engagement scheduler to exclude anyone with a booked appointment in the next two weeks. Brianna gets a courteous confirmation email instead. She replies "Looking forward to it."
Friday, May 8
Jordan sends a confirmation reminder for the upcoming Saturday-the-16th appointment. Brianna replies that night asking if she could come in earlier — she had the day off this Saturday too. Jordan moves the booking up by exactly a week.
Saturday, May 9 — Day 6
Brianna walks onto the lot at noon. Brandon, one of our sales guys, meets her at the truck. Hour and a half later, the deal is done. Truck, warranty, GAP. $4,800 in gross profit on the round trip.

What it actually proves

It's only one sale. I'm clear-eyed about that. One data point isn't a trend. But there are a handful of things this single deal proves about the way an AI BDC has to work, and they're the things I want every dealer reading this to take away.

One: speed-to-lead is the whole game, and a human team can't win it.

Brianna's lead came in Sunday afternoon. By Sunday evening she'd had a booking confirmed, a vehicle held, and a real exchange with someone she experienced as a person. That's not a luxury for an indie dealer — that's table stakes for converting an online lead at 2026 prices. The dealer who replied at 9 a.m. Monday morning lost her on Sunday at 6:14 p.m. when Jordan sent his second email.

Two: the AI being good is necessary but not sufficient. The plumbing has to be good too.

Brianna's booking exposed two real bugs the same night it was created. If either had landed in front of her, the deal probably doesn't happen. The first would have double-booked her test drive slot. The second would have sent her a weird, robot-sounding follow-up at 1 p.m. the next day that broke the spell. Both got fixed because I was watching live — and because LotLink isn't a chatbot bolted into a CRM. It's a working BDC running on plumbing I own end to end.

The dealer who replied at 9 a.m. Monday lost her on Sunday at 6:14 p.m. when Jordan sent his second email.

Three: the close is still human work. That doesn't change.

Brandon closed this deal. Jordan got Brianna to the appointment. Brandon read the room, walked the truck, navigated the warranty conversation, ran the numbers on financing, and earned the gross. That handoff — AI brings the customer to the door, human takes it home — is the whole architecture of what I'm building. Anybody selling a dealership on "AI that closes deals end-to-end" is selling them a problem. The good AI knows where to stop.

Four: confirmation emails are not throwaway.

The single most important Jordan email in this whole sequence wasn't the first one. It was the Friday-night confirmation that prompted Brianna to write back "actually, can I come earlier?" Half the dealers I know don't send appointment confirmations at all. The other half send a SaaS-template one with a barcode that nobody reads. The Jordan version is warm, specific, mentions her truck by name, and treats her like a customer. That email pulled the close in by a week.

What this cost us

Round trip — from Brianna's first form submission to keys in her hand — Jordan generated:

Forty cents. Four thousand eight hundred dollars in gross profit. That ratio is not going to be the average — I'm calling this out so I don't get accused of cherry-picking later. The average will involve more touches, more dropped threads, and more deals that just don't happen. But the BDC tier we're selling to dealers right now starts at $499 a month. One Brianna a month makes Jordan free for the year.

What's next

I'm onboarding LotLink for the next dealership in our pipeline now. Same Jordan, same playbook, slightly different voice and team-routing config tuned to their lot. Brianna's sale is the proof point I needed to walk into that conversation with my head up.

If you're an indie dealer in Northern Indiana — Warsaw, Plymouth, Goshen, Mishawaka, Fort Wayne, anywhere along the corridor — and you've watched Sunday-night leads die before Monday morning more times than you can count, that's the problem we built this to solve. The pricing page has the math. Or call me at 260-229-9393 — I answer my own phone.

To Brianna: thank you for the trust, for showing up early, and for being the first one. The truck looks great on you.


A note on attribution: Jordan handled the first-touch, the appointment booking, all follow-ups, and the confirmation that pulled the appointment forward by a week. Brandon closed the deal in person at House of Carz on May 9, 2026. House of Carz is LotLink's beta dealership and is owned by the author. Story shared with the customer's written consent.

Run the Jordan playbook at your own lot.

Indie dealers in Northern Indiana — same setup we ran at House of Carz, tuned to your team and inventory. Three tiers, transparent pricing, no annual contracts. Or call me directly and I'll walk you through the math live.

← Back to all posts