David Solar, Founder of Liminal, built his AI Show and Tell presentation around a gap most dealerships already sense but rarely quantify: a customer data platform holds far more opportunity than any team has time to find manually.

Solar presented at the AutoIndustry.AI Summit 2026, walking through the platform live in a sandbox environment built to mirror a real dealership's data.

Your DMS, your CRM, your phone system, chat, inventory, and every bit of website behavior all flow into one place. Liminal's system alone tracks around 700 different signals from a single website visit.

Once you organize that data into rules, you get a segment — a specific group of customers who share something in common. Some of these are easy to spot:

You build these, you run them, they work. But every dealer Solar's team works with, whether it's a single store or a 70-location group, runs into the same wall afterward: thousands of smaller opportunities stay buried in the data, since nobody has the time to dig them up by hand.

This is where Liminal's AI, nicknamed Les (short for Live Engagement Sequencing System), comes in. Les sits there constantly watching the data roll in and flags opportunities your team would probably never spot on their own.

During the demo, you could watch it happen in real time — someone submits a trade-in estimate while looking at a RAV4, someone else requests a test drive on a Hyundai Palisade, a payment calculator gets used, a vehicle page gets viewed. All of it happening at once.

From that stream, Les starts suggesting groups worth reaching out to: payment shoppers, people showing high engagement while still sitting outside your lead list, shoppers eyeing similar vehicles, aging inventory matches. Click into any suggestion and you'll see how many people are in it, why Les thinks it matters, and what kind of outreach makes sense. One click, and that segment goes live for one-to-one outreach across email, direct mail, and text.

The best part? Les keeps everything updated on its own. People get added the second they qualify and dropped the second they fall out of the criteria, so you're always working off a genuinely live list that stays fresh week to week.

Here's the part that got everyone's attention. Solar just typed a question straight into the system: "List everyone who bought an F-150 from us 3 to 4 years ago." Les came back with 247 people, noted that 89% of them had valid contact info, and even gave the average purchase price.

Then he just kept going, one question after another:

Then he said, "let's reach out to these 38 people," and just like that, a real, ready-to-go segment showed up inside the CDP. A plain conversation turned into an activated campaign in about a minute.

Someone in the crowd asked the obvious question: where do you even start with all this?

Solar's answer was refreshingly simple — figure out what you actually want to do before you worry about the data itself.

Once you know your use case, you can work backward and figure out what data actually matters. Transaction and service history usually come first. CRM and lead data depend on the store. And once you know what you want in there, Liminal runs it all through automatic cleanup — hygiene, normalization, NCOA — before any of it touches the AI.

TL:DR: Garbage in, garbage out applies here just like everywhere else.