A welcome reception, a full foundation day, and a hands-on build day. Here's exactly how the time runs — from the first handshake to the closing keynote.
Every seat for September 16–18 is claimed. Join the waitlist and you're first in line if a seat opens — and first to know when the next cohort is announced.
Cancellations happen. Waitlist members get first right of refusal, in order.
Join the Waitlist →1 helper for every 12 attendees · agenda subject to refinement · Group discounts available for 3+ attendees
Build your own dealership dashboard in 90 minutes. We'll rebuild the Revenue Recovery screen on your data: the one that finds lead sources that used to produce sales and quietly stopped, and puts a dollar figure on winning them back.
You bring one CSV out of your CRM. You leave with a working, deployed, password-protected dashboard on your own numbers, in a repo you own.
The point of the session is that there's no black box. Every headline number, closing ratio, average gross, recoverable revenue, is calculated from a plain monthly export. You'll see exactly how, and you'll be able to build the next screen yourself.
You won't be writing code. You'll be describing what you want, reviewing what Claude Code builds, and catching the numbers that look wrong, which is the actual skill worth having.
period, store, source, lead_type, leads, sales, new_sales, used_sales, gross, budget. In most CRMs this is "Lead Source ROI" or "Lead Source Performance". No customer PII — aggregate counts only.Can't get the export in time? Come anyway. The workshop repo ships a realistic 36-month sample dataset. You'll build the whole dashboard and swap your own file in afterwards. Nobody sits out.
Your dashboard answers the questions you already knew to ask. The conversation finds the ones you didn't.
Dashboards run on connections. Once those connections exist, the dashboard is only one thing you can do with them. This session is the other three: asking questions nobody built a tile for, setting an agent to run overnight and speak up only when a number crosses a line you set, and reconciling two systems that hold the same fact and report different numbers. That last one is where the money is.
You leave with: your own morning brief, built to your seat.
https://www.autotrader.com/car-dealers/longview-wa/70231/bud-clary-chevroletTwo formats cover almost every video your store needs. We build both, and we build them so they do not look like AI made them.
If you hear "AI video" and think marketing, it's time to think bigger. Nearly every video a dealership needs is one of two shapes: a product video where the vehicle is the hero and nobody is on camera, and a presenter video where a real person delivers information. Learn those two and you can point them at any department. HR onboarding is a presenter video. F&I product value is a product video. The service pay-over-time explainer, the thank-you video, the walkaround on your VDP, all the same two shapes.
We build a vehicle video off a real listing and an acquisition video for the customer thinking about selling you their car. Then the part that matters: the five rules separating a usable video from slop, and the critique loop where Claude watches your video back, tells you what is wrong with it, and you regenerate against that. Slop with your logo on it is worse than no video at all.
You leave with: one finished vertical video built from your own store's inventory, both production prompts ready to reuse, and a locked brand setup so the next one matches.
In this hands-on session you'll build a design system in two layers: the OEM standards you're locked into, and the store identity that's yours. Build it once and Claude applies it to everything after.
From there you'll use that system to produce a model comparison page, a marketing campaign, an inventory analysis, and a monthly one-pager performance report. One system, four very different documents.
The foundation is what compounds. A system built once should hold its shape across a customer landing page, an inventory analysis, and a report your executives read. That's what you're building first, and it's why the rest goes so fast.
You leave with a design system and a full set of working assets, not slides.
Group up with your colleagues and friends and build. This is the hour the whole workshop turns into one open floor: your laptop, your store's real problems, and every presenter in the building within arm's reach.
Gray, Dustin, Brian, Daniel, Elijah, Lisa, Alex and our partners are all on the floor. Stuck on a prompt? Wiring up a connector? Halfway through the dashboard you started this morning? Flag someone down. Think of it as study hall, except the tutors build dealership AI for a living.
Come with the thing your store is missing. Start it here, with help three feet away, instead of alone at your desk next Tuesday.
You leave with a real build in motion, not a to-do list.
Show us what you're most excited about, or something you've built that's flat-out killing it. Three days of building comes down to this: your work, on the screen, in front of a room that actually gets it.
Bring the agent that saves your BDC an hour a day. The dashboard your GM actually opens. The video nobody can tell was AI. Polished or held together with duct tape, if it works, show it.
This is your time to show off your AI prowess. Nobody in this room calls that bragging. We call it proof.
The best ideas in the room leave in everyone's laptop.
From the company behind the #1 AI agent for dealerships: turn scattered DMS data into a morning command center and build custom AI tools live.
You'll need: your work email address, and admin-level access to a Google Analytics property for one of your stores. Not the admin? Get through step 7 below, then email the connection instructions to the person on your team who is.
Accounts are approved in advance. Launch Labs will approve all participating dealers directly before the session. Questions or issues: Garrett Roach · garrett@launchlabs.ai | Mike Morgan · mmorgan@launchlabs.ai
In this hands-on session you'll connect your GA4 to Claude using an MCP built specifically to understand dealership website data and the latest ASC (Automotive Standards Council) specification.
From there you'll stand up an AI marketing analyst that answers real questions about your site in plain English: which channels and sources drive engagement, how your campaigns and landing pages perform, where your ASC events are strong or missing.
We'll also show how FocalGraph Signal can extend that picture with first-party sale, service, and verified-household attribution for dealers who want to go deeper.
You leave with a working agent, not slides.
In this two-hour, hands-on session, you'll learn a repeatable framework for solving any dealership problem: moving from Pain → Problem → Prototype. We'll work it live, using aging inventory as our example, but you'll walk away with a method you can apply anywhere.
Pain → Problem. You already feel the pain: units aging out, maybe a hunch about why. What's missing is clarity and conviction. We'll dig into the data with Claude to understand the problem actually driving it.
With the root cause in hand, we bring in Claude's research mode for outside context, gathering trends and guidance from sources the industry trusts, and layering it with your own data. We'll work with Claude to brainstorm potential tactics and pick one to tackle first.
Prototype. Then we build it. Using the Foureyes MCP, we'll turn that one tactic into a working tool wired to your own data, ready to put to work the moment the session ends.
Once the framework holds up on inventory, we'll point it at one or two more problems, to prove it's not a one-trick exercise.
You'll leave with the framework itself, plus a working tool, built on your own data, already running when you walk out the door. No engineers required.
Foureyes is the dealer-owned data platform built for automotive. We pull together the data that normally sits in separate systems (DMS, CRM, website, and inventory), and make it usable at the rooftop and group level. The Foureyes MCP connector used in this session supplies the data and usage instructions directly to Claude.