The crowd settled back in for round two of AI Show and Tell at the AutoIndustry.ai Summit after lunch for another round of solutions and demos.

Five more companies took the stage with five more specific answers:

Calls come in, texts go unanswered, leads sit cold, and every channel runs on its own clock.

The Solution: One System, Every Channel, Centralized

Scott Traylor and Thuy Adomitis, VPs of Sales at MIA, opened with three years of results: 3 million conversations, over 300,000 booked appointments, and tens of millions of dollars generated for dealerships nationwide. MIA started as a voice-inbound BDC automation tool and expanded into outbound voice, text, and soon chat and email, all built to centralize communication across a dealership's systems.

Two live calls showed the range. An inbound call booked a Cadillac owner's engine diagnostic, pushing the correct op codes and assigning her existing service advisor automatically. An outbound call, triggered by an earlier text, booked a Subaru test drive the next morning, synced directly into the CRM as a logged activity. About 25% of MIA's total volume now runs outbound, built around a simple stat: 80% of car buyers purchase from whoever reaches them first.

Traylor also handed dealers a way to test any AI vendor on the spot: interrupt the demo mid-sentence and watch whether it forces a customer to repeat themselves, or read a long monologue and see if it knows when to respond. A vendor that fails either test is selling a demo, not a working system.

A shopper on a vehicle listing page at 9 p.m. wants a real answer immediately, not a callback once the store opens.

The Solution: A Shopping Assistant Built Into Every Listing

Dave Ferguson, who heads Spyne's US business, built his demo around a chat, voice, and email widget embedded directly on every vehicle photo across a dealership's website. The assistant carries real context: it knows a shopper's engagement history, pulls live vehicle history data through AutoCheck and Carfax integrations, and answers detailed questions on the spot.

A live call showed a customer asking about a 2026 Mercedes-Benz GLC 300, cargo space for a 50-pound dog, dealership hours, and the store's exact address, all answered instantly in one conversation. Spyne also runs outbound campaigns tied to CRM data, reaching an entire database with one consistent, well-timed message when an incentive like a $1,000 loyalty offer drops.

Missed calls during the morning rush, declined services nobody follows up on, BDC conversations that stall right before booking. None of it shows up as one clear number, but all of it adds up.

The Solution: Containment, Coaching, and Automated Follow-Up

Corbin Witte, GTM at Avoca, built his fixed ops platform on a home services foundation, now running across 75% of the top 25 private-equity-backed operators in that industry. Avoca's call tree pulls a customer's DMS record in real time and books routine service work cleanly, while a human-in-the-loop team of 80 trained agents handles the 8-12% of calls that need real empathy, like billing disputes.

Avoca's Coach feature scores every call against a dealership's own rubric. One example: a customer called about a brake job, the BDC agent stayed polite, and Coach still flagged the call at 54 out of 100 for never confirming urgency or offering a specific time slot. A fixed ops director got an alert within 15 minutes, called back, and booked the appointment.

Avoca's Nurture campaigns target declined services directly from the DMS, timed for the two-to-six-week window when a customer has had time to reconsider without forgetting the conversation. One home services client recovered $800,000 in a single season using the same approach. Witte closed with three questions dealers can ask any AI vendor: does it book natively into your DMS, what's the containment rate on hard calls, and how long does real onboarding actually take?

Dealerships build the obvious marketing segments and stop there, missing the smaller opportunities no human team has time to find manually.

The Solution: An AI That Finds What Gets Missed

David Solar, founder of Liminal, demoed Les, an AI engine that watches a dealership's data in real time, tracking around 700 signals from a single website visit alone. Les surfaces segment suggestions on its own, like payment shoppers or aging inventory matches, and keeps every list updated automatically as customers qualify in or out.

The sharpest moment came from plain conversation. Solar typed, "list everyone who bought an F-150 three to four years ago," and Les returned 247 people. A follow-up question narrowed that to 183 customers in a positive equity position, then 141 flagged as a service retention risk. A final request built a ready-to-launch segment of 38 people directly inside the CDP, all in about a minute of typed questions.

A new lead carries no age, no budget signal, no read on urgency, and outreach built for a stranger comes out generic by default.

The Solution: A Real-Time Coaching Assistant Built on a Full Buyer Profile

Jay Ku, founder of HeyGreenlight, built Signal to pull a full consumer profile within two seconds of a lead landing in the CRM, then feed that into Wingman, a live AI-generated coaching script a rep opens during the call itself. A live demo showed the system predicting objections, adjusting talk points in real time as a customer mentioned hauling a lawn mower and traveling with three teenagers, and pulling in a Honda-to-Hyundai comparison the moment brand loyalty came up.

Four separate agents run on every call: one listening to the customer, one to the salesperson, one feeding in real-time profile data, and one building a live DISC personality read from the conversation itself. Ku was direct about what makes this work: the underlying consumer data is available to anyone. The value sits entirely in how it gets delivered to a rep mid-conversation.