Key Takeaways:
The dealership that responds first usually wins the deal, and most stores still take too long to make that first move.
A good customer experience on a call reflects directly on the brand, whether a human or an AI answers the phone.
The dealerships getting real value from AI treat it like a new staff member that needs training, not a tool they plug in and forget.
Scott Traylor and Thuy Adomitis, VPs of Sales at Mia, opened their AI Show and Tell presentation with a detail that shapes everything else in their conversation: Traylor spent 20 years as a dealer before partnering on this technology, building Mia specifically around problems he lived through on the dealership floor.
The two presented at the AutoIndustry.AI Summit 2026, walking through three years of results, a set of live call demos, and a framework for evaluating any AI vendor a dealership might consider.
Three Years, Three Million Conversations
MIA has processed more than 3 million customer conversations since launch, and those conversations have booked over 300,000 appointments, generating tens of millions of dollars for dealerships nationwide. The platform now runs across 85 employees and more than 35 integrations.
MIA started on the sales side with a specific goal: automate the BDC function across sales, service, and reception. From there, the platform expanded from inbound voice into outbound voice, then into text, with chat and email support planned before year's end — all aimed at centralizing communication for both the dealership and the customer.
Watching the Calls in Action
Two live call recordings anchored the presentation, each showing a different side of MIA's platform.
The first call handled inbound service. A customer named Carol called in about a warning light on her 2025 Cadillac XT5. MIA identified the vehicle, offered available appointment times, and booked her in for an engine diagnostic — confirming every detail along the way. Behind that simple exchange, the system pushed the correct op codes into the appointment and assigned the booking to the same service advisor Carol had worked with over the past year, automating the administrative work a service team would otherwise handle manually.
The second call handled outbound sales. MIA had already reached a customer named Robert through an outbound text about a 2025 Subaru Cross Trek Sport. When Robert called back, MIA confirmed the vehicle's availability, price, and mileage, then booked a test drive for the following morning, syncing the result directly into the dealership's CRM as a logged activity.
About 25% of MIA's total conversation volume now runs through outbound channels, following up automatically on leads the moment they land in a dealership's CRM.
Why Speed Still Decides the Deal
Traylor grounded the outbound push in one statistic: 80% of car buyers purchase from whichever dealership reaches them first. That single number explains why automatic, immediate lead follow-up carries as much weight as any other part of the platform — a lead sitting untouched for even a few hours can already be lost to a faster-moving competitor.
How to Actually Test an AI Vendor
Traylor closed the presentation with a set of tests dealers can run on any AI demo, regardless of vendor:
Interrupt the AI mid-conversation. A well-built system handles being cut off gracefully. A poorly built one forces the customer to repeat themselves, creating exactly the kind of friction a real customer would notice immediately.
Read several paragraphs in a monologue. This exposes whether the AI can tell the difference between a pause and an actual turn to respond, a distinction that matters constantly in real conversations.
Ask how much setup work falls on you. Some AI providers lean heavily on the dealer to handle prompt engineering and ongoing configuration. Traylor's team built MIA around the opposite premise: a system that does the work itself, since most dealers came up in the business, not in engineering.
Where This Technology Goes Next
Looking ahead, MIA's team framed their roadmap around a simple structure: in a ten-step sales process, AI handles steps one, two, nine, and ten, while the middle stays with the human closing the deal. Over the next year, that means deeper involvement in lead follow-up and qualification, more advanced service-side work, and continuous self-learning that lets the system remember and refine itself from every conversation it handles.
One limitation came up directly during the Q&A: MIA works from a dealership's inventory system as its single source of truth, and a vehicle currently out on a test drive registers exactly as available as one sitting on the lot. Traylor flagged this as a genuine current limitation, noting that any AI vendor claiming real-time knowledge of a vehicle's physical location warrants real scrutiny.
The Mindset Shift That Matters Most
The biggest hurdle to getting real value from AI, in Traylor's view, has little to do with the technology itself. Dealers trained for years to treat new purchases as tools — something to install, switch on, and leave alone — carry that same instinct into AI adoption, and it works against them. MIA's team frames the right approach differently: treat an AI system like a new staff member, one that needs coaching, feedback, and ongoing development the same way any new hire would.
That framing extends to the numbers behind MIA's own growth. Three million conversations and 300,000 booked appointments came from three years of continuous refinement, not a system installed once and left untouched. The dealerships getting the most out of AI right now are the ones applying that same expectation internally, coaching the technology the way they'd coach any other member of the team.
