

The bottom line: Two different flavors of AI are showing up in the service drive right now. One has a phone bill and a revenue number attached to it. The other has a press release. Operators should know which is which before a vendor walks in asking for budget.
Here's an uncomfortable fact buried in a recent Car Dealership Guy podcast episode: service advisors at some stores are letting calls go to voicemail on purpose. Not because they're lazy — because they're slammed, and a missed call from someone wanting to check on a part or reschedule a 7:30am oil change is a call they can defer without immediate consequence. The customer usually calls back, or doesn't, and either way the advisor gets back to the sixteen other things on their desk.
That "either way" is where the money was leaking out. Voice AI platforms are now catching those calls — after-hours, during rush, during lunch — and actually booking the appointment instead of taking a message. According to the podcast, this is producing $8,000 to $30,000 a month in revenue at stores where those calls used to just die in a voicemail box. That's not a productivity metric. That's repair orders that didn't exist a year ago, on cars that were already going to need the work done somewhere.
The credibility marker worth noting: Mia is described as the first voice AI vendor approved for General Motors' co-op program. That's a real gate, not a marketing claim — GM co-op approval means someone at the OEM level looked at the product and decided it was worth subsidizing. Vendors self-report ROI numbers constantly; a manufacturer signing off on reimbursement is a different kind of evidence.
A second data point, same theme, different department: at Jim Ellis Kia, GM Mike Martucci described a three-person sales team fielding 700 leads a month — a ratio that should make any operator wince — using conversational AI from Matador AI to double close rates and push service scheduling completion to nearly 90%. The number worth sitting with isn't the close rate, which every vendor claims to double something. It's the 90% scheduling completion. Most stores, human-staffed, don't get there, because scheduling requires someone to actually answer, ask the right questions, and get the appointment on the calendar before the customer's attention wanders to a competitor's website.
The pattern across both stories: conversational AI's actual value in fixed ops isn't answering questions cleverly. It's persistence and coverage — showing up every single time the phone rings, at 9pm on a Tuesday, without needing a lunch break or a day off.
Then there's the fixed-ops "intelligence" category, exemplified by the newly announced IVSG–Chameleon Limited partnership pairing IVSG's DMS integration work with Chameleon's TimeAi and WarrantyAi platforms. The pitch: centralize service and parts data, surface trends and "threats," give leadership "total clarity in three clicks."
This is a familiar genre of dealership press release, and operators should read it with the skepticism it's earned. Fixed-ops dashboards promising unified data and hidden revenue have been pitched for over a decade — DMS vendors, third-party BI tools, and now AI-branded versions of the same idea. The concept isn't wrong; fragmented service and parts data really does hide profit leaks. But "three clicks" and "actionable intelligence" are marketing copy, not a case study. There's no dealership named, no before-and-after number, no independent account of a store that used this and found $40k a month sitting in warranty claims nobody was chasing. That doesn't mean it won't work. It means there's nothing yet to evaluate beyond the press release itself.
If you're deciding where to spend the next fixed-ops technology dollar, the evidence right now points toward the phone, not the dashboard. Voice AI for service scheduling has publishable revenue numbers, an OEM co-op approval, and a specific mechanism (catching calls humans are letting go) that explains why it works — not just that it does. The intelligence-platform category may get there, but "unifies data" and "drives accountability" are claims every fixed-ops vendor has made since the DMS was invented. Ask for a reference store before you ask for a demo.
The practical move: audit your own missed-call and voicemail rate in service before buying anything. If it's high, you already know where the $8,000 a month is hiding — it's the call your advisor let ring out this afternoon.

ACV Auctions and DriveCentric are linking service-drive inspections directly to trade-in sourcing and fixed-ops upsell, while Numa's acquisition of SocketTime aims to sync customer-facing AI with real-time shop capacity. Both moves point the same direction: stop treating service and sales as separate systems that don't talk to each other.

Numa, the AI operating system already texting and calling customers for 1,300+ dealerships, has bought fixed-ops software company SocketTime to get inside the actual service scheduling, dispatch, and inspection workflow. The pitch: stop bolting AI onto the front counter and start wiring it into repair orders, technician capacity, and parts approvals. It's a smart bet on paper — but the integration, promised by year-end, is where these things usually stall.