Voice AI for Automotive
Dealerships lose an average of $1.17M per year from missed service calls. Not from bad marketing or slow technicians, but from calls that ring out, go to voicemail, or sit on hold until the customer books somewhere else.
A voice AI agent answers every inbound call 24/7, books service appointments directly into your DMS, and routes sales inquiries to whoever's available on the floor. No hold queue. No voicemail pile. No separate system your advisors have to reconcile the next morning.
Silverthread Labs builds these with native DMS integrations: CDK Global, Reynolds & Reynolds, DealerSocket. Appointment data goes into your existing system, not into a silo someone has to sync later.
the revenue sitting in your voicemail#
42% of service calls go unanswered, at $450 per repair order#
The average dealership misses 158 bookable service calls per month, according to Numa's 2024 Industry Trends Report. At $450 per repair order, that's roughly $71,000 in recoverable service revenue per month, per rooftop. Annualized: $1.17M.
That's a call handling problem, not a conversion rate problem or a technician capacity problem. Most dealer principals already know this. What's harder to pin down is whether a fix that doesn't require hiring more service advisors actually exists.
sales leads sitting on hold while floor staff are with customers#
Sales inquiry calls during peak floor traffic, a Friday afternoon or Saturday morning, frequently go unanswered or sit on hold long enough that the prospect hangs up and books elsewhere. A single lost vehicle sale is $2,000 to $4,000 in front-end gross for a typical franchised store. If your phones are a bottleneck during your busiest hours, you're leaving that gross on the table consistently.
why peak hours (8-11:30 AM, Monday through Tuesday) are where most losses happen#
Numa's 2024 data puts Monday through Tuesday, 8:00 AM to 11:30 AM, as the highest-miss window at the average dealership. The service drive is packed with write-ups, the phones are ringing for new appointments and same-day status calls, and service advisors are physically unable to pick up.
The callers who get dropped in that window typically planned ahead. They called first thing Monday to get scheduled, which means they were ready to book. Those leads don't usually come back.
what voice AI does in an automotive context#
service appointment booking: caller intent, availability check, and DMS write-back#
A voice agent handles the full service booking flow: it identifies the caller's need (oil change, brake inspection, warranty concern), checks available appointment slots from your DMS scheduler in real time, confirms the time with the caller, and writes the appointment record directly back into CDK, Reynolds & Reynolds, or DealerSocket.
No separate calendar. No separate system for the advisor to check in the morning. The appointment shows up in your existing scheduler the same way a human-booked call would.
Dealers using AI-assisted inbound service agents convert 56% of booking opportunities into confirmed appointments, with top-performing stores reaching 70% conversion, according to STELLA Automotive AI's 2025 data.
sales inquiry routing: qualify, collect contact info, transfer or schedule callback#
For sales calls, the agent captures the caller's vehicle of interest, trade-in situation, and preferred contact method, then transfers to an available salesperson or schedules a callback. If nobody is available, the lead record writes into your CRM (DealerSocket, VinSolutions, or another system) rather than disappearing into voicemail.
The agent does not try to close a vehicle deal. Its job is to get every inbound sales inquiry to a human, with the caller's information already captured, rather than leaving that person to call back cold.
parts availability inquiries and hours/location handling#
Parts calls and directional calls are a meaningful share of inbound volume and require no human judgment to handle. A voice agent answers these accurately and consistently, which frees service advisors to focus on customers already in the drive.
after-hours coverage: take bookings when the service desk is closed#
The service desk closes at 6 PM. Customer availability to schedule service does not. An after-hours voice agent takes appointment requests during closed hours, writes them into the DMS scheduler for the following morning, and sends confirmation to the caller. When the service desk opens, the overnight bookings are already there.
DMS integrations: where the work actually happens#
CDK Global: API-based appointment creation and scheduler access#
Our integration uses CDK's API layer to read available appointment slots from your configured scheduler and create appointment records directly in CDK, including service type, technician assignment if pre-configured, and customer record matching against existing CDK profiles.
Reynolds & Reynolds: ERA-IGNITE integration for service writing#
Our agents integrate with ERA-IGNITE to book service appointments, retrieve customer vehicle history for returning callers, and write new appointment records without requiring manual re-entry by service staff.
DealerSocket: CRM and scheduling sync for leads and appointments#
DealerSocket covers both the CRM function (for sales lead capture) and the scheduling function (for service). Our integration handles both paths: sales inquiry leads write into DealerSocket CRM with caller-captured data, and service appointment requests book directly into the scheduling module.
what DMS-connected means vs. a standalone booking system#
A standalone booking system creates appointment records in its own database and then either syncs to your DMS on a schedule or requires manual transfer. Both create reconciliation work and raise the odds of double-booking.
DMS-connected means the agent writes directly to your source of truth. When the service desk opens, the overnight and overflow bookings are already in the scheduler your advisors use, formatted the same way a human-booked call would be. For dealer groups with non-standard DMS configurations, multi-brand rooftops, or routing logic that doesn't fit a packaged product, that distinction matters more than it sounds.
how we build it#
step 1: audit current call handling and DMS configuration#
Before writing a single conversation flow, we audit your inbound call volume, DMS scheduler configuration, and existing telephony setup. We document the escalation paths that require a human, because not everything should go to the agent. This step produces a scoped build specification: what the agent handles, what it escalates, and how it connects to your DMS.
step 2: build conversation flows for service, sales, and parts#
Separate conversation flows for each call type, with branching logic that matches how your service team actually talks to customers. Your service director or GM reviews and approves the flows before testing starts.
step 3: integrate telephony (Twilio or Telnyx) and DMS write-back#
Telephony runs on Twilio or Telnyx depending on your existing setup. We configure the agent on your existing service line (no number changes required in most cases) and wire up DMS write-back through the appropriate API or integration layer.
step 4: test with live call routing before cutover#
Before going live on your main service line, the agent runs on a portion of routed calls. Results get reviewed against what a human advisor would have done. Testing continues until accuracy and booking rates hit the targets defined in the build spec.
step 5: monitor, tune, and hand off with documentation#
Post-launch, we track call handling metrics, booking conversion rates, and escalation patterns for 30 days. Conversation flows that show drop-off get tuned. At the end of the monitoring period, you get full documentation: conversation flows, integration configuration, escalation logic. Your team can run the system from there without us.
Pricing#
single-rooftop setup#
Single-rooftop engagements cover the full build: telephony configuration, conversation flow design and testing, DMS integration, and 30-day post-launch monitoring. Pricing varies by DMS integration complexity and the number of conversation flows.
Most engagements fall in the range of one missed month of service revenue. The payback math is straightforward.
Request a scoped estimate for your rooftop
multi-location and dealer group pricing#
Dealer groups with multiple rooftops on the same DMS can share integration infrastructure, built once and deployed across locations, with location-specific conversation flows on top. Group pricing starts at three locations. If rooftops run different DMS platforms, we scope each integration separately within a consolidated project.
Contact us to scope a dealer group deployment
FAQ#
How much revenue do car dealerships lose from missed calls?
About $1.17M per year in recoverable service revenue, based on 158 missed bookable calls per month at $450 per repair order (Numa 2024 Industry Trends Report). That number covers service only; it doesn't include lost sales leads.
Can voice AI integrate with CDK Global or Reynolds & Reynolds?
Yes. We've built DMS integrations for CDK Global (API-based appointment creation), Reynolds & Reynolds ERA-IGNITE (service writing), and DealerSocket (CRM and scheduling). Appointments write directly into your scheduler. No parallel system, no manual reconciliation.
How does an AI agent book service appointments at a dealership?
The agent identifies the caller's service need, checks available slots from your DMS scheduler in real time, confirms a time with the caller, and writes the record directly into your DMS. The booking shows up in your existing scheduler exactly as a human-booked call would.
What does it cost to set up a voice AI agent for a dealership?
It varies by DMS complexity and the number of conversation flows. Most single-rooftop builds fall around one month of recoverable missed service revenue. Contact us and we'll give you a scoped estimate based on your actual setup.
How does voice AI handle sales inquiry routing at a dealership?
The agent captures vehicle of interest, trade-in details, and preferred contact method, then transfers to an available salesperson or schedules a callback. The lead writes into your CRM so the salesperson has context before the call. The agent doesn't try to close the deal.
What happens when the agent can't handle a call?
Every conversation flow has defined escalation points where the agent hands off to a live person or schedules a human callback. We design and review that logic with your team during the build phase.
How long does implementation take?
Most single-rooftop builds take four to six weeks from initial audit to live deployment. The main variable is DMS configuration complexity and how long your service director needs to review the conversation flows before testing starts.
If missed calls are a real problem at your service desk, it's worth finding out what fixing it actually involves. Request a call audit and we'll scope what a build looks like for your specific DMS and call volume.
