Auto Glass Business Automation in 2026: Triage, Part Lookup, Insurance, and ADAS

Auto glass is a parts business wearing a service business costume

Most trades win or lose on labor and scheduling. Auto glass wins or loses on whether the right piece of glass is on the van when the technician pulls into the driveway. Everything upstream of that — the phone call, the intake questions, the insurance conversation, the calendar — exists to make that one thing true.

That makes the auto glass intake call unusually high-stakes. A homeowner calling a plumber can be vague and the plumber still shows up useful. A vehicle owner who says "I need a windshield for my Ford" has given you almost nothing: that same model year can carry four or five distinct part numbers depending on whether it has rain sensing, a lane camera, acoustic interlayer, heating elements in the wiper park area, or a head-up display reflective layer. Guess wrong and you have burned a technician's morning, a customer's patience, and a restocking fee. As of August 2026, that remains the most expensive recurring mistake in the trade, and it is almost always traceable to a question nobody asked on the phone.

The auto glass companies growing without adding office staff are the ones that turned the intake interview into a system instead of an act of memory. This guide covers the whole operations stack: answering, damage triage, part identification, insurance versus cash, mobile versus in-shop scheduling, ADAS recalibration, weather, billing, and reviews. It maps to the Run with Jarvis platform.

Why auto glass loses more calls than it should

The technician is the phone. In most auto glass shops under fifteen people, the person who knows which glass fits is the same person cutting out urethane. A caller who reaches voicemail at 9:40 AM is a caller who is already dialing the next result.

Demand is event-driven and shops as a batch. A rock hits a windshield on the highway and the driver calls three companies from the shoulder within ten minutes. There is no research phase. The first company that answers, sounds like it knows the vehicle, and gives a date usually gets the job. The decay curve here is steeper than in almost any other trade, which is the same dynamic covered in the speed-to-lead guide.

Hailstorms produce volume that no single person can absorb. One severe storm produces a week of calls compressed into two days, all of them for the same three or four vehicle platforms. A human answering sequentially loses most of that spike. The trigger is a weather event rather than a calendar month, but the shape is the same as any seasonal spike.

Insurance callers hang up on hold. A customer trying to use their glass coverage is already braced for friction. If your phone makes them wait, they will call the number on the back of their insurance card instead, and the carrier will route them to a network shop that is not you.

Layer 1: An AI receptionist that knows the glass questions

The foundation is KeyBot, the AI phone agent. It answers on the first ring in English or Spanish, at any hour, and runs your intake script the same way on the third call of a hail week as on the first.

Two properties matter more here than in most trades.

Parallel answering. Auto glass demand arrives in bursts tied to weather and highway conditions. Eight simultaneous calls after a storm is normal. A person handles one. The AI holds all eight, and every caller gets a real conversation instead of a hold queue.

Script fidelity on a long interview. Auto glass intake is genuinely long — vehicle identification, feature identification, damage description, insurance status, location, access. A human doing that interview thirty times in a day starts skipping the questions that seem redundant, and the skipped question is always the one that changes the part number. Automated intake asks all of it every time, which is the entire reason downstream part ordering gets more accurate.

A traditional answering service is the wrong tool here, because it takes a message and hangs up without any of the fields you actually need. A phone menu is worse: someone standing on a highway shoulder is not pressing three.

Layer 2: Triage — repair or replace, decided on the phone

Chip repair and full replacement are different products with different prices, durations, insurance treatments, and margins. Deciding which one a caller needs is a verbal triage that automates well because the deciding inputs are all things the customer can see and describe.

The questions that actually separate the two:

• How large is the damage, compared to a coin the customer has in the car? A small bullseye or star break is repair territory. A long crack is not.
• Is it in the driver's primary viewing area? Repairs leave optical distortion, and distortion directly in the driver's sightline is a reason to replace rather than repair even when the damage is small.
• Has the crack reached the edge of the glass? Edge cracks compromise the structural bond and generally mean replacement.
• Did it grow after the defroster ran or after a cold night? A spreading crack is a replacement conversation, and telling the customer that up front prevents a wasted repair visit.
• How long ago did it happen? Contamination in an old chip changes the cosmetic result, which changes what you should promise.

Getting this right on the phone has an outsized effect on both revenue and reputation. A repair booked as a replacement is an unnecessarily expensive job the customer may cancel after a second opinion. A replacement booked as a repair is a technician driving out with resin and a UV lamp for a job that needs a full glass set, which is a lost slot and an apology.

The triage should also route to the right calendar block. A chip repair is a short appointment that can be slotted between larger jobs, often at a customer's workplace parking lot. A replacement is a longer job with cure time attached. Treating both as generic appointments is how a route gets destroyed.

Layer 3: Pinning the exact part — VIN and features

This is where auto glass intake diverges from every other trade, and where structured automation pays for itself outright.

Year, make, and model are the opening question, not the answer. The features that change the glass:

• Rain sensor — needs a specific bracket and gel pad arrangement.
• Forward-facing camera for lane keeping and emergency braking — changes the bracket, and triggers recalibration.
• Acoustic or laminated interlayer — a quieter cabin glass that customers notice immediately if you substitute the cheaper version.
• Heated windshield or heated wiper park area — requires an electrical connection the standard part does not have.
• Head-up display — needs the reflective wedge layer, and a non-HUD substitution produces a ghosted double image that will come back as a warranty call.
• Solar or infrared coating and shade band variations — visible differences the customer will compare against the other side of the vehicle.
• Antenna and defogger elements in back glass, and whether a quarter glass or vent glass is encapsulated.

The reliable way to resolve all of this is the VIN, which is why VIN capture belongs in the intake script rather than in a callback. The AI asks the customer to read the VIN from the base of the windshield on the driver's side or from their insurance card and registration, and confirms it back digit by digit. Every one of those characters lands in the customer record before anyone touches a parts catalog.

Ask for photos as part of the same call. A follow-up text asking for a photo of the damage, a photo of the mirror-area sensor housing from inside the vehicle, and a photo of the VIN plate resolves nearly every remaining ambiguity. Texting is the right channel for this because the customer is already holding the phone.

The record then lives in IntelliDrive, so the vehicle, the VIN, the feature list, the part ordered, and the outcome are all attached to the customer. The second time that fleet or that family calls, none of it has to be rebuilt. IntelliDrive is the CRM, dispatch, and invoicing layer for exactly this reason.

Layer 4: Insurance versus cash, sorted at intake

These are two different businesses sharing one phone number, and mixing them up costs money in both directions.

Cash and out-of-pocket callers want a number. They are price shopping, often against two other companies, and a vague answer loses them. Your intake needs to produce a real out-the-door figure that already includes the glass, the moulding and clips if the vehicle needs new ones, mobile service if applicable, and recalibration if the vehicle has a camera. A quote that later grows by a recalibration fee is the single most common source of driveway arguments in this trade.

Insurance callers want reassurance and a short path. What you need from them is the carrier, the policy or claim number if one exists, whether a claim has already been filed, the deductible situation for glass specifically, and whether their state or policy treats glass differently from other comprehensive claims. What they need from you is a clear statement of what happens next and who calls whom.

The AI captures all of that verbatim, which matters because a garbled policy number turns into a rejected submission and a week of delay. The intake also flags the deductible conversation for a human when the caller's answers are ambiguous, rather than improvising coverage advice.

Fleet is a third path. Commercial accounts have purchase orders, approval chains, and vehicles the driver knows nothing about. Fleet calls should be recognized at intake and routed to the account record instead of being interviewed like a retail customer.

Layer 5: Mobile versus in-shop, cure time, and the calendar

Auto glass scheduling has constraints most service calendars do not model well.

Cure and safe drive-away time is real and it belongs on the appointment. Urethane needs time before the vehicle is structurally sound, and that window varies with product and conditions. If your calendar treats a windshield replacement as an install duration with no attached cure window, you will hand a customer keys too early or block a bay for a duration nobody wrote down.

Some jobs cannot be mobile even when the glass could be. Static recalibration requires level floor, controlled lighting, target boards, and clearance in front of the vehicle. That is a bay, not a driveway. The intake has to know which vehicles require static procedures so the job is booked in-shop from the start, rather than discovered by a technician who has already driven forty minutes.

Mobile jobs carry site conditions that determine feasibility. Covered parking or a garage, a level surface, whether the vehicle can sit undisturbed during cure, whether the customer will be present, gate codes, apartment complex parking, and whether there is shade. Those fields belong on the appointment so the technician is not calling the office from the curb.

GetTimePad holds real availability so the AI books during the call rather than promising a callback, and it holds those site fields with the appointment. The general mechanics are in how AI appointment booking works. Customers who would rather book online at midnight land in the same calendar, so a mobile slot and a bay slot never get double-sold.

Layer 6: ADAS recalibration as a first-class billable step

Camera-equipped vehicles have quietly turned windshield replacement into a two-part job, and shops that still treat recalibration as a favor are giving away both revenue and liability protection.

Operationally, recalibration needs its own treatment:

• Its own duration and resource. A static procedure occupies a bay and a target setup. A dynamic procedure occupies a technician and a stretch of suitable road with clear lane markings, which is a scheduling constraint tied to daylight and traffic.
• Its own line item on the quote. If it is not quoted at intake, it becomes a surprise at the driveway, and surprises at the driveway become chargebacks.
• Its own documentation. A completed recalibration record attached to the job is the thing that protects you if the vehicle is later in a collision. Dispatch notes, timestamps, and photos all belong on the job record.
• Its own sublet path when you do not do it in house. If you send camera vehicles to a partner, the handoff is a second appointment with a second location, and the customer needs to be told that at booking rather than at pickup.

The simplest rule: any vehicle whose intake shows a forward-facing camera should generate a two-step job automatically. That is a workflow rule, not a memory exercise, and it is the kind of thing you can express in plain language to the Jarvis brain rather than building a custom form for it.

Layer 7: Weather, reschedules, and the day that falls apart

Mobile auto glass is weather-dependent in a way that in-shop work is not. Rain, freezing temperatures, and high humidity all affect adhesion and cure. A wet Tuesday morning does not delay one job; it invalidates a route.

Rebuilding that day manually is an hour of calling performed by whoever should be pulling glass. Half the customers do not answer, so it becomes voicemail tag that leaks into Wednesday.

Handled as an automation, the affected appointments are flagged, the customers get a text with the reason and real open slots from the live calendar, and the rebooking writes itself back into the schedule. Rescheduled appointments should also carry their own reminder sequence, because a reschedule has a higher no-show rate than an original booking — the customer's urgency cooled with the weather.

The same flag matters in the other direction: a caller who reaches you after hours during a storm should still get a booked appointment rather than a promise.

What each plan gives an auto glass company

Every plan is month-to-month with zero setup fees, unlimited users, and no per-call fees.

Running the arithmetic on your own numbers

Use your own ticket, not an example ticket.

Assume a shop completes 70 jobs a month at a $420 blended average across chip repairs, standard replacements, and camera vehicles with recalibration. That is $29,400. Assume the phone rings 210 times a month and, because the person who knows the glass is also cutting it out, 60 of those calls go unanswered.

Assume only 7 of those 60 would have closed. At $420 that is $2,940 in recovered monthly revenue. Core at $500, plus roughly 340 overage minutes at $0.45 for the extra volume, is about $653. The recovered revenue is more than four times the cost before counting anything the scheduling, recalibration tracking, or billing layers contribute.

The minute math matters in this trade because the calls are long. A full auto glass intake — vehicle, features, VIN readback, damage triage, insurance, location — realistically runs five to seven minutes, longer than a pressure washing or lawn care call. At six minutes, Core's 500 included minutes cover roughly 83 calls. A shop taking 210 calls a month is looking at something like 760 overage minutes at $0.45, about $342, for a $842 effective monthly cost. Once you are consistently past about 1,100 minutes, Pro's 1,000 minutes and $0.40 overage is the cheaper structure, and it adds attribution on top. Hail months push a lot of shops over that line for two or three months a year, which is a good argument for choosing a plan on your peak rather than your average. The framework is in the AI receptionist ROI guide.

Where attribution starts earning its keep

Auto glass is one of the most competitively advertised local categories there is, and raw call volume is a badly misleading success metric. A campaign can triple your ring count by attracting people asking whether you buy scrap glass or whether you fix a cracked phone screen.

Pro adds Dynamic Number Insertion, so each call ties back to the campaign, keyword, and ad that produced it. Conversion data feeds back to Google Ads so bidding optimizes toward booked replacements rather than toward ringing phones. Transcription and lead scoring tell you whether a keyword produces camera-equipped replacement jobs or chip repairs, which is a large difference in ticket value. The general mechanics are in the call tracking and attribution guide.

Two adjacent automotive stacks are worth reading side by side with this one, because the intake structures overlap: the mobile mechanic automation guide for the mobile-service scheduling patterns, and the auto repair shop automation stack for the bay-and-estimate side.

Reviews and the local search funnel

Almost nobody has a preexisting relationship with an auto glass company. They search, they scan ratings, they call the top few. That makes review volume close to the entire top of your funnel.

Review automation asks for the review right after a completed install, while the customer is looking through clean glass and the technician just cleaned up their dash. That timing is the whole trick, and it never happens reliably when it depends on a person remembering. The approach is in how to get more customer reviews.

For general small business planning material, the Small Business Administration publishes guidance at sba.gov. For anything touching recorded calls, review the Federal Trade Commission's consumer protection material at ftc.gov alongside your state's requirements.

A sensible rollout order

Answering and intake first. That is where the recoverable revenue is, and where the part-number accuracy comes from. A structured VIN and feature interview on every call changes your ordering error rate immediately.

Then booking, with mobile and bay treated as different resources, and cure time modeled on the appointment.

Then dispatch, arrival texts, and the recalibration second step, so camera vehicles stop being handled by memory.

Then invoicing and payment links, so the technician collects before leaving rather than mailing an invoice. That mechanic is covered in getting paid faster with payment links.

Then review automation.

Only after all of that, if you are spending on advertising, move up for attribution and campaign management. Measuring ad performance precisely while still missing sixty calls a month only tells you exactly how much you are losing.

The short version

Auto glass is decided by who answers, who identifies the right part on the first call, who quotes the recalibration up front instead of at the driveway, and who keeps a weather day from collapsing.

Answer every call in parallel. Triage repair against replacement with the same five questions every time. Capture the VIN and the feature list before anyone opens a catalog. Sort insurance from cash at intake. Book mobile and bay work as different resources with cure time attached. Make recalibration a scheduled, quoted, documented second step. Automate the weather reschedule.

If you want that modeled against your own call volume, ticket mix, and camera-vehicle percentage, reach out or review the plans on the pricing page.