Service Winback
Your best service customers are quietly leaving. Win them back and prove it.
AutoRelay finds the customers who stopped coming in, re-engages them with a consent-gated AI conversation that answers real objections, books them a real appointment, and measures the results against a control group, so the number you see is incremental, not inflated.
The problem
You’re losing service customers earlier than you think
The industry’s own research says the service relationship starts leaking during the warranty period, while an aging car parc pushes more repair demand into the market every year. Their numbers, not ours:
Franchised dealers’ share of U.S. service visits, 2018 → 2025
Source: Cox Automotive, 2025 ↗New-vehicle owners still servicing with their selling dealer, 2023 → 2025
Source: Cox Automotive, 2025 ↗How it works
Find, reach, book, and prove it
Find them
The lapsed list builds itself from data you already have: no visit in 6–12+ months, declined work on file, or an open recall. Opt-outs and anyone already booked are excluded automatically.
Reach them
An identified AI agent texts each customer and handles the real objections: price questions without invented numbers, and “no time” with actual open slots.
Book them
It reads the exact day and time back, gets an explicit yes, and books it, 24/7, straight onto your schedule, badged as AI.
Prove it
A control group is never messaged. Your report counts only the extra return visits, matched to real repair orders at 30, 60, and 90 days.
Playbooks
Ready-made playbooks, one click to launch
No prompt-writing, no list-building. Every playbook starts in review mode: nothing sends until you approve it.
Lapsed Customer Re-engagement
Customers who haven’t been in for 6–12+ months. A warm check-in on the vehicle, answers to their questions, and an easy booking.
Declined Services Follow-up
Customers who declined specific work on a past visit. The opener names the exact operation and the visit date, and the agent gets it scheduled without ever quoting a stale price over text.
First-Visit Follow-up
Customers who came exactly once and never returned: the biggest loyalty leak. The softest touch: a relationship check-in first, a booking second.
Open Recall Outreach
Safety-first outreach for vehicles matching an open recall. The store confirms applicability and parts before anything is promised, and the repair itself is free from the manufacturer.
Always-on
Not a campaign someone remembers to run
Turn on a rule and winback runs itself: when a customer crosses a trigger, they’re enrolled in the right playbook automatically. You set the daily pace; the safety rails are built in and always on.
The rails are hard-coded: no consent, no contact. Already talking to them? Skipped. Upcoming appointment? Skipped. Contacted recently? Skipped. And every skipped customer is logged with the reason, an audit trail you can read, not a black box.
Measured, not claimed
Every campaign ships with its own control group
A slice of every audience is held out and never messaged. Your report compares the two groups and counts only the extra customers who returned for paid service, matched to real repair orders at 30, 60, and 90 days. If a campaign doesn’t beat its control group, the report says so. Small control groups are flagged. Estimated dollars are never blended with matched dollars.
The gap between the bars is the only number we report: incremental return visits.
Illustrative example of the method, not a claimed result.
And the report cuts the answer the way you’ll be asked for it: by playbook, by customer tenure, and by how fast customers came back after the message.
That’s the number your GM, and anyone auditing your marketing spend, actually wants: incremental repair orders, not response rates.
Compliance
Built consent-first, because outreach to past customers has rules
Consent-gated audiences
Promotional winback only goes to customers with a stored SMS-marketing consent record, a gate that configuration cannot switch off. Strict-consent mode verifies against your consent provider and fails closed.
STOP honored everywhere
Opt-outs are enforced at the application level across every channel, on top of carrier handling: immediately, mid-conversation, permanently.
Quiet hours & caps
No automated sends outside local business hours. Volume caps per campaign build, one active winback conversation per customer, and cooldowns between contacts.
Humans stay in charge
Campaigns start in review mode. Staff see every AI message and booking, badged as AI, and can take over any conversation at any time. Every AI action is logged and auditable.
Pilot results, coming soon
We publish measured, holdout-verified results, not marketing math. Dealer pilot case studies will appear here as they complete, with the store’s permission.
FAQ
Service Winback FAQ
How is this different from service marketing we already run?
Most service marketing is broadcast (ads, mail, or batch email) and reports responses. AutoRelay reaches out 1:1 over SMS from your own customer data, holds a real conversation that ends in a booked appointment, and reports incremental return visits against a never-messaged control group, matched to actual repair orders.
What does the AI never do?
It never quotes prices or discounts, never diagnoses a vehicle over text, sends at most one gentle follow-up, and hands complaints or anything safety-related to your team. It books only after reading the exact day and time back and getting an explicit yes. These rules are enforced by automated tests, not just prompts.
How do you handle consent and opt-outs?
Winback audiences are consent-gated at build time: promotional outreach only goes to customers with a stored SMS-marketing consent record, and that gate cannot be switched off. STOP is honored across every channel on top of carrier-level handling, quiet hours are enforced in your local timezone, and every new campaign starts in review mode so nothing sends until you approve it.
What data do you need from us?
Whatever is already syncing: scheduler and DMS data feeds cover it. No DMS write access is required: bookings land on AutoRelay’s scheduler with full staff visibility and automated reminders, and every booking is verifiable against your appointment data.
Is this a campaign I have to remember to run?
Either. Playbooks run as scheduled campaigns, or you turn on always-on rules: when a customer crosses a trigger, such as lapsing past your window or matching an open recall, they’re enrolled in the right playbook automatically, within a daily cap you set. The safety exclusions are hard-coded, every skipped customer is logged with the reason, and each rule reserves its own control group so always-on results are measured exactly like campaigns.
Why do you hold customers out of the campaign?
A small control group (10% by default, adjustable) is how we prove the campaign caused the return visits instead of claiming credit for customers who would have come back anyway. The group stays stable across re-runs, small samples are flagged, and if a campaign underperforms the control group, the report says so.
See your own lapsed list before anything sends
The audience preview shows exactly who would be reached, who’s excluded and why, and the size of your control group. Book a demo and see it on your data.
Also from AutoRelay: service-lane vehicle acquisition.