01. The Problem
Three numbers that should keep you up at night.
An independent auto repair shop sits on a goldmine of work it never bills. The work is already approved, the customer already trusts you, the vehicle already needs it. You just never asked the second time. Here is how this playbook models three of the most common revenue leaks in the bay.
This is built on an illustrative model for independent auto repair shops. The model spans single-bay weekend operators to 12-bay full-service outfits doing alignments, diagnostics, brake jobs, timing belts, and fluid services. Its figures are modeling assumptions, not measured results. Swap in your own shop data wherever you have it.
Three leaks. Each one quantifiable. Each one targeted by a module in this playbook.
Each of these is a system failure, not a customer failure. Customers don't track their own brake pad thickness. They don't remember which tech recommended the timing belt last visit. They don't flag the registration sticker on the windshield. That work belongs to the shop. If your shop management software (Tekmetric, Shop-Ware, Mitchell1, NAPA TRACS, AutoVitals) is already storing the data, the only missing link is the outreach layer that turns the data into a booked RO.
The five modules in this playbook are built to close that gap. None of them require a new shop management system, a new POS, or a new database. They sit on top of what you already have and they reach out to the customer at the moment that matters most.
02. Outcomes
What the model targets for a shop on the Nirvani stack.
These are illustrative targets, not measured results. They come from this playbook's model of a shop 90+ days after launch, not from a cohort of real shops. Module 01 (AI quote agent) is built to move conversion first. Modules 02 and 05 take the longest to compound.
03. The Stack
Five modules. Each one fixes a leak. Each one stacks.
You can deploy any single module first. You can deploy all five in week one. The recommended sequence is the order below: start with the AI quote agent (Module 01) because it is built to stop new leakage first. Then bolt on the declined-service follow-up (Module 02), which lights up your existing pipeline of missed revenue. Modules 03, 04, and 05 are the compounders that build a return-customer flywheel.
The number one leak in a service-writer-driven shop is the call that comes in while the writer is standing next to a lift explaining a brake job to another customer. The phone rings. It goes to voicemail. The caller hangs up. The caller calls the next shop on the Google search results. The work goes to your competitor on Main Street.
The AI quote agent answers immediately. It is trained on your shop's MAP labor pricing, your common labor times for the top 40 jobs (brake pad replacement, oil change, alignment, AC recharge, timing belt, coolant flush, transmission service, battery, alternator, starter, water pump), and your part-ordering rules (OEM vs aftermarket preference, in-stock vs special-order). It quotes a range, books a drop-off slot, and sends the customer a confirmation SMS.
For jobs that require diagnostics (a check-engine light, an OBD2 code pull, anything outside the common-job library) the agent books a diagnostic appointment and tells the customer the diagnostic fee upfront. It never quotes a number it cannot defend.
- Always-on phone coverage. Inbound calls answered 24/7, even when the lobby is closed.
- Real labor time lookups. Pulls from Mitchell1 or your shop management system's labor time database.
- Bay schedule visibility. Books into open slots, respects lift availability and tech-hours load.
- ASE-certified tech routing. Routes diagnostics to your senior tech, oil changes to the apprentice bay.
- SMS confirmation. Customer gets a text with date, time, address, and what to bring (registration, prior service records).
That sample call runs about 47 seconds. The agent gave a price range, booked the slot, and prompted for the upsell ("anything else while it's on the lift"). The ROI model below assumes that question adds $74 per RO in ticket size, an illustrative assumption, not a measured result.
Every shop runs multipoint inspections. Most shops write the declines into the RO ("customer declined timing belt, advised due at 90k") and then forget about it. The customer leaves. Six months pass. The timing belt actually fails. The customer is on the side of I-95 calling AAA. The shop that recommended the work first does not get the tow-in revenue because nobody followed up.
The declined-service follow-up sequence reads the structured decline data out of your shop management software (Tekmetric, Shop-Ware, Mitchell1, NAPA TRACS, AutoVitals all expose this) and triggers a follow-up cadence at the exact recommended-by date for each service type.
- Timing belt declined at 90k. SMS at 88k miles, second nudge at 95k, voice call at 100k.
- Brake pads at 30% (recommend by 6 months). SMS at month 5, voice call at month 7.
- Coolant flush declined. Reminder at 4-month anniversary of the recommendation.
- DEF system service declined (diesel). Three-month follow-up with current state-of-fluid warning language.
- Alignment declined post-tire purchase. SMS at month 2 with "uneven wear pattern" warning.
Hi Linda, Maple Street Auto. Last visit we noticed your 2018 CR-V's front pads were at 30%. They'll likely need replacement in the next 8 to 12 weeks. Want us to hold a slot? Reply YES or pick a time at maplestreetauto.com/book. Reply STOP to opt out.
The sequence runs unattended. The customer does not need to call in. The work-order conversation happens entirely over text, a channel customers check fast: in EZ Texting's 2026 survey of 959 US consumers, 87% said they check a new text within 15 minutes. For high-ticket declined work (timing belts, head gaskets, transmission service over $1,200) the system upgrades to a voice call from the AI agent because the spoken word converts higher on larger tickets.
In states that require periodic safety inspections or emissions testing (rules vary by state and county, and Texas ended most passenger-vehicle safety inspections on January 1, 2025), the inspection deadline is the single most predictable recurring service event. Every car needs one per year. Every customer forgets when theirs is due. Whoever reminds the customer first wins the inspection and any failed-component repair that comes with it.
The reminder system pulls the inspection-due date from the customer's vehicle record (stored on the registration tab in your shop management system or scraped from the sticker date on the prior inspection invoice) and schedules a three-touch sequence ending the week before expiration.
- 30 days out. SMS: "Your 2018 CR-V is due for state inspection by April 15. Book now."
- 14 days out. SMS with calendar link and price ($25 inspection, $7 emissions per local statute).
- 7 days out. AI voice agent calls to book. Modeled at ~38% conversion on this call (an illustrative assumption).
That call surfaced an upsell that would have otherwise gone to whoever ran the inspection. The AI agent is trained to ask the "noises or lights" question on every inspection booking. Illustrative upsell assumption: +$112 per inspection RO (a modeling assumption, not a measured average).
Every vehicle has an OEM-recommended maintenance schedule. The 2018 Honda CR-V needs an oil change every 7,500 miles, a tire rotation every 7,500 miles, a transmission fluid service at 60k, a spark plug service at 105k, and a timing chain inspection at 120k. The OEM publishes the schedule. Your shop management software (Mitchell1, AutoVitals, NAPA TRACS) has the schedule loaded. The customer has no idea any of it exists.
The maintenance schedule engine takes the customer's VIN, pulls the OEM maintenance schedule, tracks the customer's actual mileage (from the last RO odometer reading plus an average-miles-per-day extrapolation), and surfaces the next due service in the form of a personalized SMS.
- VIN decode on first visit. Decodes year/make/model/trim/engine/transmission.
- OEM maintenance schedule mapping. Pulls Honda Maintenance Minder, Toyota Care, Ford recommended service, etc.
- Mileage extrapolation. Tracks days since last visit + miles delta to estimate current odometer.
- Personalized SMS. "Your 2018 CR-V is due for a transmission fluid service in the next 1,500 miles."
- Book-now CTA. One-tap link to the bay schedule, pre-loaded with the recommended service.
Hi Linda, Maple Street Auto. Based on your 2018 CR-V's mileage, your transmission fluid service is due in the next 1,500 miles. Honda recommends this at 60k to extend transmission life. It's a 90-minute service, $189. Book a slot at maplestreetauto.com/book. Reply STOP to opt out.
The agent did not pressure. It cited the OEM source. It offered the schedule for the customer to verify. It still got the booking, because the customer trusted the answer. That is how maintenance-schedule outreach should always feel. Pushy gets blocked. Educational gets booked.
A strong moment to ask for a Google review is about 90 minutes after the customer picks up the vehicle and drives it. The brakes feel firm. The alignment feels straight. The lights work. The customer is feeling the value. That is the moment to ask. Not three days later. Not at the counter when the customer is in a hurry to leave.
The review acceleration module fires an SMS 90 minutes after the RO is closed in your shop management software. The SMS asks the customer to rate the visit one to five, and every reply gets the same Google review link. One, two, or three stars also alerts the shop owner, so the owner can call and make it right. Google's review policy bars selectively asking only happy customers.
- 90-minute trigger after RO closed. Tied to your shop management system's "vehicle delivered" event.
- Same link for every rating. Every customer, 1 to 5 stars, gets the Google review link. No gating.
- Pre-filled Google deep link. Opens the Google review form with the shop pre-selected.
- Owner alert on low ratings. Negative feedback triggers a phone call from the owner within 4 hours.
- Service-writer attribution. Each review tagged with the writer's name for performance tracking.
Hi Linda, hope the CR-V is driving better. On a scale of 1 to 5, how was your visit today with Maple Street Auto? Reply with just a number. Reply STOP to opt out.
Compare the 1-star branch:
The owner alert is what turns an unhappy visit into a personal phone call from the owner, the best shot at winning that customer back, while every customer still gets the same review link. The model targets +1.4 Google reviews per week with a star average moving from 4.3 to 4.7 inside the first 90 days (illustrative targets, not measured results).
04. Implementation
30 / 60 / 90 day roadmap.
You do not deploy all five modules on day one. You sequence them so each one compounds the next. Module 01 is built to add new conversion from week one. Module 02 reaches into your back catalog of declined services. Modules 03, 04, and 05 build the return-customer flywheel behind the model's 2.4x visit-frequency target over the next quarter.
- Deploy AI Quote-to-Approval Agent (Module 01).
- Connect to Tekmetric / Shop-Ware / Mitchell1 / NAPA TRACS via API.
- Load top 40 common jobs with MAP labor pricing.
- Train agent on shop tone, service-writer style, technician routing rules.
- Go live with after-hours coverage first, then daytime fallback.
- Track quote-to-book conversion week-over-week.
- Activate Declined-Service Follow-Up (Module 02).
- Backfill the last 12 months of declined ROs into the sequence.
- Activate State Inspection / Emissions Reminder (Module 03).
- Build the 30/14/7-day inspection cadence for every active customer record.
- Monitor declined-service recapture rate. Target 12%+ by day 60.
- Tune the cadence pacing based on customer response signals.
- Deploy Maintenance Schedule Engine (Module 04).
- VIN decode every customer record. Map OEM schedules.
- Activate Review Acceleration (Module 05).
- Connect 90-minute SMS trigger to RO-closed event.
- Run weekly metric review with shop owner: conversion, recapture, frequency, reviews.
- Lock in the playbook. The stack runs without daily attention.
Implementation is handled by the Nirvani onboarding team. The shop owner spends roughly 6 to 9 hours total across the 90 days. That includes the initial intake call, the script review for Module 01 (we recommend service-writer voice and tone), and three weekly check-in calls during the first month. After day 30 the stack runs on its own with monthly check-ins for tuning.
The shop does not need to change shop management software. Tekmetric, Shop-Ware, Mitchell1 Manager SE, NAPA TRACS, AutoVitals, Identifix, ALLDATA Manage, and Auto-Repair-Cloud are all supported via API. If your software is not on the list, we offer a manual import workflow that runs on a nightly batch.
05. ROI Snapshot
What the math looks like for a 3-bay independent.
An illustrative 3-bay shop running 22 ROs per week with a $480 average ticket. Pre-stack baseline. Modeled post-stack at day 90. Every input below is an illustrative assumption, not a measured result.
Illustrative math. Pre-stack revenue: 22 ROs/week x $480 x 4.3 weeks = $45,408/month. Modeled post-stack revenue: 31 ROs/week x $554 x 4.3 weeks = $73,848/month. Delta: $28,440/month in modeled incremental revenue. Typical engagement investment is scoped on the strategy call. In the model, the first incremental RO lands in week 2.
For multi-location operators or shops above $1M annual revenue running multi-channel marketing, a multi-location engagement adds the dedicated voice number cluster, the Yelp / Google Business Profile syndication layer, and the advanced customer-segment routing that lets you treat a fleet customer differently from a retail customer.
06. Illustrative Example
3-bay independent. 90 days. Modeled numbers.
Maple Street Auto. Suburban metro market. 3 lifts, 1 alignment rack.
Maple Street Auto is a fictional shop built to show the math. It is not a real customer, and every figure in this example is modeled. It is the kind of shop that runs on word-of-mouth and Google. Owner Mike has been turning wrenches for 28 years. His service writer Jamie has been there 11 years. They run on Mitchell1 Manager SE. The shop does brake jobs, timing belts, AC work, alignments, state inspections, and the occasional engine swap. Average ticket is $480. Pre-deployment monthly revenue: $45,408.
Mike was losing calls every day. The phone would ring during a customer counter conversation. Jamie would let it go to voicemail. Maybe 1 in 5 voicemails got returned same-day. The rest went to the next shop down the road. Mike knew it was happening but he could not justify hiring a second service writer at $52k/year fully loaded for what felt like a phone-coverage problem.
In this example, Mike deploys the Nirvani auto repair stack on a Monday. Day 1 the AI quote agent answers calls. By day 14 the declined-service follow-up sequence is processing 11 months of backfilled declined ROs. By day 60 the inspection reminder cadence is live. By day 90 the maintenance schedule engine is personalizing outbound SMS to every customer record. Here is what the model projects at day 90.
In the model, the 44% conversion lift comes almost entirely from after-hours and during-counter-conversation calls that would otherwise go to voicemail. The $26,471 in recovered declined revenue comes from timing belts (8 ROs at an average of $1,840), brake services (24 ROs at $340), and fluid services (19 ROs at $189). The star rating climb comes from 47 net new Google reviews in 90 days, of which 41 are 5-star, 4 are 4-star, and 2 are 3-star (both flagged to Mike by the low-rating alert, both turned into regulars after his personal call).
The point of the example is the mechanism, not the figures: the declined work was already sitting in Mike's Mitchell records, waiting for someone to follow up.
Get the full auto repair playbook by SMS and email.
We will send you a 12-page PDF with the deployment-ready scripts for all 5 modules, the Tekmetric / Shop-Ware / Mitchell1 connection guides, the MAP labor pricing template, and a 30-minute strategy slot with the founder if you want one.
On the way.
Check your inbox for the 12-page PDF and your phone for a single confirmation SMS. Sammy will reach out within 24 hours if you flagged the strategy-call option.
Frequently Asked Questions
Does this work with my shop management software?
If you are on a less common system, the manual import workflow runs a nightly CSV batch that handles all five modules with a 24-hour lag instead of real-time. We model effectiveness roughly 12 percentage points lower because of the lag (an illustrative assumption, not a measured result).
How does the OBD2 integration work? Do we need a scanner per bay?
For shops that want code-level pre-qualification on inbound calls, we offer an optional customer-facing OBD2 reader workflow where the agent walks the caller through a free phone app that reads the code over Bluetooth and texts it back to the shop. This is a Tier 2 feature reserved for shops on N2 or higher.
How do parts ordering and OEM vs aftermarket pricing flow through the AI quote?
Parts ordering itself happens on your shop's existing parts platform (NAPA, AutoZone Pro, O'Reilly First Call, WorldPac SpeedDial, AAIA Aces, RockAuto for niche cases). The agent does not place the order. It books the appointment and the tech writes the parts order at the start of the RO based on the diagnostic and customer-approved parts type.
Can the AI agent dispatch a technician or move work between bays?
Mid-day re-routing (when the senior tech finishes early, when the alignment rack frees up, etc.) stays with your service writer. The agent surfaces the open slot, the writer makes the call.
What happens when a customer needs a tow? Can the agent dispatch one?
The agent does not negotiate tow rates or guarantee tow availability. That stays with the tow operator. The agent's job is to keep the customer's vehicle in your network instead of letting them call a random tow company that delivers to whichever shop pays the kickback.
How is the SMS A2P 10DLC compliance handled? We've had carriers block our texts before.
Carrier filtering (the blocks you have experienced before) usually happens because the sender is using an unregistered campaign or shared shortcode. The Nirvani auto repair stack uses dedicated long codes registered to your shop's business identity. The Campaign Registry says any business texting customers on 10DLC must register.
What does this cost and is there a contract?
Month-to-month, no annual commitment. Cancel any time. The payback target for the illustrative shop in this playbook is about 6 weeks, a modeled figure, not a measured result. If your shop has not reached payback inside 90 days, we will personally keep tuning the deployment with you.