The Declined and Deferred Repairs Your Shop Is Quietly Losing Every Week

The declined work sitting in your shop management system right now

Your lead tech pulls a 2015 Silverado onto the lift for a brake job. Fifteen minutes into the digital vehicle inspection he flags a weeping water pump, a torn CV boot, and lower control arm bushings that are finished. The service advisor sends the inspection with photos to the customer through Tekmetric. The customer approves the brakes and declines the rest: not today, maybe next month. Your system files three line items under declined work, worth roughly $1,400 in parts and labor, and nobody opens that report again.

By Thursday the same advisor has four estimates over $1,000 waiting on approval. One is an $1,800 suspension job. The customer said he would call back after talking to his wife, so the car sits on your lift tying up a bay. Two days later he authorizes the work at the dealer down the road, because they called him and you did not. Friday at 8:50pm a driver with a dead alternator calls your shop, gets voicemail, and books somewhere else before you open.

None of this shows up as a crisis. There is no alert and no red number. It leaks quietly: a few declined jobs here, an estimate that goes cold there, an after-hours call that rings out. Across a busy multi-bay shop that is the gap between an average repair order around $480 and one closer to $620, and between four cars a day and six. Tekmetric, Shopmonkey and Mitchell 1 record all of it accurately. None of them lifts a finger to win it back.

Problem 1: The declined jobs nobody ever chases

Every shop running digital inspections is building a graveyard of declined work. The report exists in Tekmetric and Shopmonkey. What does not exist is anyone with time to work it. Your advisors are slammed writing today's orders, and last month's declined water pump is not their concern at 4pm on a Friday.

The off-the-shelf tools stop at the report. They can show you the declined line items, but they will not text the customer 45 days later, reference the exact vehicle and service, attach the original inspection photo, and offer a time. A custom follow-up engine does exactly that. It reads declined and deferred line items from your system through its API, sequences a message at a sensible interval (sooner for a leaking water pump, before winter for a weak battery), and personalizes it to the vehicle and the specific job. Replies route straight to the assigned advisor, and every booked recovery is tracked back to the line item so you can finally see the revenue the report was sitting on.

Problem 2: Estimates that go cold waiting on approval

The $1,800 suspension job did not lose itself. It sat unapproved for two days while your advisor was heads-down and the customer was undecided. Tekmetric sent the estimate once. After that, silence, until the dealer got there first.

A shop management system treats an estimate as a document. It does not know that an unapproved job over $1,000 is money walking out the door. Custom automation treats approval as a workflow with a clock on it. When a high-value estimate goes unapproved past a threshold you set, the AI sends a plain, friendly nudge with the inspection photos, answers the predictable questions (why this is needed now, what happens if it waits), and offers a financing link if you carry one. If the customer replies with a real objection or a question about price, it hands off to the advisor with the full context instead of dropping the ball. Shops measure this directly as a lift in inspection approval rate.

Problem 3: The 8pm call that goes straight to voicemail

Plenty of the cars that book on a Tuesday were decided the night before, and your phone was off. The stranded alternator caller, the parent whose brakes are grinding on the school run, the fleet manager with a truck down: they call after five, hit voicemail, and the next shop's phone rings.

Tekmetric, Shopmonkey and Mitchell 1 do not answer your phone. That is not what they are for. An AI phone and booking agent does. It picks up after-hours and overflow calls, tells the caller your hours and whether you service their vehicle, captures the year, make, model and complaint, and books the appointment straight into your calendar against real bay and technician availability so it does not double-book your one alignment rack. It texts a confirmation, logs the call, and escalates a genuine emergency to a human. The outcome the owner cares about is simple: the calls that used to become voicemails become cars in bays.

Problem 4: The reviews you earned but never asked for

You did clean work on that Silverado. The customer drove off happy. No review went up, because asking got forgotten in the rush of pickup, and the one time you did ask, it was at drop-off before the work was even done.

Some systems fire a generic review request, usually mistimed and easy to ignore. A custom flow triggers on the moment that matters, the repair order closing and the car leaving, and sends a personalized ask by text a short time after pickup. It routes an unhappy reply to the owner privately before it lands on Google, and it tracks which advisor and which technician are earning the reviews. Wired to your Google Business Profile, this is the difference between a trickle of reviews and a steady stream that quietly moves you up the local map pack.

Problem 5: The schedule that double-books your best tech

Your master tech is the only one who touches diagnostics and driveability. This morning the calendar has him on two big jobs at once, a comeback squeezed in between, and a car pulled into the bay before its parts arrived. Nobody planned it that way. The calendar simply does not know any better.

A shop management calendar is a grid of open slots. It does not understand the labor hours on the estimate, which tech is certified for the job, which bay has the lift or the alignment rack, or whether the part is on the shelf. Smarter scheduling built around your shop does. It reads the job's labor time, the technician's skills, the bay type, and parts availability, and it stops a car from being pulled in before the parts land. For shops running fleet accounts or a mobile service van, the same logic routes the mobile tech so you are not sending him back across town twice in one day.

Problem 6: The gold mine already inside your shop management system

You have years of repair orders, customers and vehicles sitting in Tekmetric or Mitchell 1, and almost none of it has ever been worked as a marketing asset. The customer who came once in 2023 for a battery and vanished. The fleet of company vans overdue for their interval by mileage. The two hundred vehicles with declined work you already know about.

That database is not a filing cabinet, it is a demand engine nobody switched on. Mining it means pulling service history, mileage and intervals, then automatically reaching the customers who are actually due: an oil change coming up by date or estimated mileage, a timing belt approaching its window, a one-time customer worth a win-back offer. This is not a blast to your whole list. It is the right message to the right owner about their specific vehicle, which is exactly the thing a generic email tool inside your management system was never built to do.

What this costs and how long it takes

These bands come from Digital Heroes delivery experience across more than 2,000 projects, not a price list. A focused first release, the kind that solves one expensive problem, typically runs $50,000 to $120,000 and ships in 10 to 16 weeks. For a shop, that is usually the declined-work and approval follow-up engine, or the after-hours booking agent, wired into Tekmetric or Shopmonkey. A full operations platform that also handles scheduling, reviews, reactivation and multi-location reporting runs $150,000 to $350,000, phased over 6 to 12 months so you get working pieces along the way instead of one distant launch.

What pushes the number up in this niche is specific. Some shop systems have solid APIs and some do not: Tekmetric and Shopmonkey expose real integration points, while an older Mitchell 1 Manager SE install can mean deeper work to read and write data cleanly. Texting your customers legally means A2P 10DLC registration and TCPA-compliant opt-in, which is not optional and takes lead time. Parts and labor guide integrations, financing and payment processing, and rolling the whole thing across several locations each add scope. None of that is a reason to avoid building. It is the reason to scope it with someone who has done it in a shop before.

When Tekmetric is enough, and when it is time to build

Be honest with yourself before you spend a dollar. If you run a single location with one or two advisors, a standard workflow, and you are actually using the declined-jobs report and the built-in texting, Tekmetric, Shopmonkey or Mitchell 1 is enough. Do not build software to solve a discipline problem.

The signals that it is time to layer AI automation on top are concrete. You have multiple locations with inconsistent processes. Your service manager burns two hours a day chasing approvals and declined work by hand. There are thousands of declined line items rotting in the system. You carry fleet or wholesale accounts with billing your software fights you on. The phone goes to voicemail every night while you are closed. The position that has held up across 2,000-plus builds is this: do not rip out Tekmetric, it is good at what it does. Put an AI layer on top of it through its API, and only replace the core system in the rare case where the system itself is the thing blocking your shop.

How to choose a developer for auto repair software

The wrong developer will build you a pretty generic CRM (Customer Relationship Management) that does not know what a declined job is. Vet for the specifics of this trade.

First, ask what they have integrated with. Someone who has worked with Tekmetric, Shopmonkey or Mitchell 1 can tell you honestly which systems have real APIs and which need a workaround, before you sign. Second, make sure they understand A2P 10DLC and TCPA texting rules; a shop that starts blasting texts without registration gets its numbers blocked, and preventing that is the developer's job. Third, listen for whether they speak your economics: average repair order, effective labor rate, inspection approval rate, comebacks, declined work. If they only say "CRM" and "leads," keep looking. Fourth, insist they prove it against a copy of your real repair order history before you commit to a large build, so you see it working on your Silverados and your customers, not a blank demo.