AI for Auto Repair

Turn a repair order, estimate, or inspection note into a structured ADAS-calibration documentation packet — identifying which Advanced Driver Assistance Systems are likely to require calibration based on the work performed, citing the OEM position statements and service procedures that trigger the requirement, and producing a clean paper trail the shop can attach to the RO for insurance billing, liability protection, and customer communication.

Investigate a single comeback — a vehicle returning with the same complaint, a related complaint, or an unintended consequence of recent work — and produce a structured root-cause analysis that distinguishes between technician execution, parts quality, diagnostic miss, sublet vendor failure, and customer-side factors. Then aggregate patterns across the shop's RO history to identify systemic comeback drivers, and produce both a per-comeback corrective action plan and a shop-level prevention recommendation.

Turn a technician's spoken narration — captured on a phone, headset, or shop-tablet voice recorder — into a structured, advisor-readable Repair Order (RO) note that lives where the rest of the job lives: on the RO itself. The output reads like the tech wrote it carefully, not like a raw transcript. Cause / Correction / Customer-facing language are separated, measurements are flagged for accuracy, parts and labor operations are pulled out so the advisor can grab them at a glance, and the next-shift handoff is explicit.

Build a personalized, week-by-week ASE certification study plan for a shop technician or apprentice — covering which topics to study, in what order, how many hours per week, which free and paid practice resources to use, and what to focus on in the final week before the exam.

Score a real service-advisor phone call against a 5-dimension rubric, identify the moment-of-inflection turn that decided the outcome, surface missed upsells and weak objection handling, and produce a one-thing-to-rehearse-next-time coaching note the manager can use in a 1:1. Input is a call transcript (and optional metadata like booked / not-booked, RO total, advisor name); output is a rubric score with quote anchors, a missed-opportunities list, a customer-sentiment read, and a single coaching focus. Different from `service-advisor-roleplay-coach.md` (which rehearses a hypothetical against an AI customer before the call) and from `service-advisor-script.md` (which produces the live script the advisor reads off of). This is the after-the-fact audit layer — looking at what actually happened on a real call, with consent, to make the next call better.

Guide technicians through a structured, step-by-step diagnostic process based on DTC codes, reported symptoms, and vehicle history — producing a ranked list of probable causes with the fastest confirmation tests for each.

Turn a technician's raw inspection notes, measurements, and photo references into a structured, customer-ready Digital Vehicle Inspection (DVI) report — with a red/yellow/green severity schema, plain-language findings, specific measurements, photo captions, recommended action tiers, and a prioritized work plan. The output is formatted so an advisor can send it via SMS/email link, walk a customer through it at pickup, or hand off directly to the `service-advisor-script.md` skill for live presentation.

Translate a technician's symptom description or verified diagnosis into a clean, counter-ready parts order worksheet — with correct part name, OE and aftermarket part number patterns, fitment warnings for trim/engine/drive variations, supersession flags, a related-parts-usually-replaced-together list, a vendor sourcing plan mapped to the shop's preferred suppliers, and the labor-op code that will feed the Repair Estimate Builder. The output is formatted so the counter person (or tech acting as counter) can pull quotes in two browser tabs without re-entering vehicle data.

Run a structured roleplay rehearsal between a service advisor and an AI-played customer persona, then score the transcript against a shop-defined rubric and produce a coaching note the manager can use in a 1:1 review. The advisor practices a specific scenario — a price-reveal conversation, an objection rebuttal, a do-not-drive safety conversation, a declined-work walk-through, a high-emotion warranty escalation — against an AI customer who pushes back realistically. After the practice run, the AI scores the interaction on a 5-dimension rubric, surfaces specific advisor quotes that worked or didn't, and produces a brief coaching note. Different from `service-advisor-script.md` (which produces the live script the advisor reads off of) and from `technician-onboarding-sop-generator.md` (which produces SOPs the team follows). This is the rehearsal layer — practice reps, low stakes, before the real customer.

Turn the tribal knowledge locked in a senior technician's head into a written, repeatable Standard Operating Procedure that a new hire or apprentice can follow without shoulder-surfing. The input is a 5–20 minute verbal walkthrough (transcript or bullet notes) from the experienced tech. The output is a finished SOP with a tool-and-parts checklist, numbered steps, safety callouts, quality-control checkpoints, a sign-off block, and a small "common mistakes" appendix — formatted so it can be printed and taped to a toolbox or dropped into the shop's knowledge base.

Turn deferred repairs and declined recommendations into booked appointments by generating personalized, non-pushy follow-up messages that reference the specific work the customer passed on, the safety or cost implications of waiting, and a clear, low-friction path back into the shop.

Turn verified diagnostic findings and a technician's job plan into a clean, itemized, customer-ready repair estimate — with parts, labor, sublet, shop supplies, tax, disclaimers, warranty terms, and optional tiered pricing (required / recommended / deferred). The output is structured so the advisor can paste it into the shop management system (Tekmetric, Shop-Ware, AutoLeap, Mitchell1, Protractor, R.O. Writer) or send it as a standalone PDF/SMS link for customer authorization.

Turn a customer's open declined work, scheduled preventive-maintenance intervals, and known wear items into a single forward-looking 12-month vehicle care plan — a month-by-month roadmap that sequences the work by safety priority, spreads cost across the year, sets clear booking windows, and gives the customer something to keep on the fridge or in their glovebox. Output is the document that gets handed to the customer at pickup or emailed the next morning, plus a shop-side summary that hands cleanly into `declined-services-followup.md` for the message campaigns and into `maintenance-reminder-sequence.md` for the per-interval nudges. Different from a follow-up message (which is a single nudge) and different from a one-time interval reminder (which is a single service line) — this is the annualized plan that the customer references for the next twelve months.

Generate the full system prompt, conversation-flow tree, and escalation rules for an AI voice or SMS receptionist that answers every inbound call to the shop 24/7 — capturing appointment requests, qualifying emergencies, collecting vehicle and customer details, and handing off cleanly to a human advisor when the situation requires judgment.

Translate raw tire-inspection data — from AI vision tools like Anyline TireBuddy, TyreSwift, UVeye, or a manual tread-gauge + visual check — into a clear, jargon-free customer explanation that tells the vehicle owner exactly what each corner of their car is showing, how urgent it is, and whether they need new tires, an alignment, both, or just a watch-and-rotate. The output is built for a service-advisor walkthrough at pickup, an attached PDF summary, or an SMS/email to a customer who already left the shop.

Produce parallel English + Spanish versions of any shop-to-customer message — appointment confirmations, status updates, estimate approvals, declined-work follow-ups, pickup-ready notices, review requests, or bad-news calls about hidden damage — in tone, terminology, and register that a native US-Spanish-speaking customer would actually use. The output is two messages the advisor can send to the right customer without re-translating anything and without sounding like a machine.

Translate raw battery health test results — from tools like BATTSCAN, Dr.EV, AVILOO, GreenTech, or a manual multi-meter cell test — into a clear, jargon-free customer explanation that tells the hybrid or EV owner what their battery is doing right now, how urgent any issues are, and exactly what they should do next.

Turn a single open RO (or a batch of open ROs for the same fleet account) into the complete B2B communications package the shop needs to run that account cleanly: an approval request to the fleet's maintenance decision-maker (with a clear recommended action and the cost delta against the account's authorization threshold), status updates routed to the right fleet contact (not the driver), a consolidated multi-vehicle weekly report, and an exception / escalation note for anything outside the account's routine-approval rules. Fleet and B2B customers do not want the same voice a retail customer gets — they want terse, structured, priority-classified communication they can act on in 30 seconds between meetings.

Generate clear, friendly customer-facing text or email messages for each stage of a repair job — from vehicle drop-off through completion and pickup — so service advisors never have to compose status updates from scratch and customers always know where their vehicle stands.

Generate a personalized, multi-touch reminder sequence (initial, follow-up, final) for scheduled maintenance — based on standard mileage intervals, seasonal services, and customer's last visit — ready to deploy via SMS, email, or both.

Write clear, calm, and respectful customer messages when parts costs change between estimate and invoice — whether because of tariff pass-through, a vendor price increase, a superseded part number, a backorder substitution, or an OEM-vs.-aftermarket swap. Produce three artifacts the shop can use immediately: a phone-call talk track for the service advisor, an SMS / email re-quote message for the customer, and a revised estimate summary line the advisor can paste into the shop management system.

Generate thoughtful, on-brand responses to Google, Yelp, and Facebook reviews of the auto repair shop — whether 5-star praise or 1-star complaints — so the owner or service advisor can reply to every review within 24 hours without writing each one from scratch.

Turn an open NHTSA or OEM safety recall (or a list of open recalls across the shop's active customer database) into a complete, compliance-safe multi-touchpoint outreach sequence: an initial SMS + email on Day 0, a voicemail drop on Day 1–2, a follow-up text at Day 4, a follow-up email at Day 7, a final text at Day 14, and a monthly nudge thereafter until the recall is closed or the customer explicitly opts out. For "do-not-drive" critical recalls, the sequence compresses to a daily-contact cadence until the vehicle is scheduled at an authorized dealer. Output is tuned for independent-shop voice (not dealer voice) — the shop is the trusted messenger who flags the recall, explains it plainly, routes the customer to the OEM-authorized dealer for the recall completion itself, and stays in the relationship for all the non-recall work the customer will still need.

Draft the exact words a service advisor should use when presenting diagnostic findings and recommended work to a customer — structured as a live phone or in-person script, not a written estimate. The output covers the opening, the findings walkthrough, the recommendation tiers (required / recommended / deferred), the price reveal, the most common objections, and the explicit close. Designed to raise authorization rates without overselling, and to give junior advisors a repeatable framework that matches the shop's voice.

Organize all available repair documentation into a structured, submission-ready chargeback rebuttal packet — including an evidence narrative, supporting-documents checklist, and a merchant dispute response letter — so the shop can respond to credit-card chargebacks without hiring outside help or missing the bank's 7–30 day response window.

Turn a teardown finding (hidden damage revealed after disassembly of a visibly damaged area) plus the original estimate plus a set of timestamped teardown photos into a clean, insurer-ready supplement request: line items with correct operation descriptions, labor times cited to the published labor guide (CCC, Mitchell, Audatex), parts cited with source and price, a cause-and-effect narrative tying the impact pattern to the hidden damage, an OEM-position-statement citation line for every structural or safety-system component, a photo index keyed to each line item, and a carrier-specific formatting variant (State Farm Select Service, GEICO ARX, Progressive DRP, Allstate Good Hands, Nationwide Blue Ribbon, Farmers Circle of Dependability, USAA Preferred, or non-DRP). The output is designed to reduce the back-and-forth round-trips that kill cycle time, by anticipating the specific line-level questions an adjuster will ask and answering them in the supplement itself.

Audit a batch of closed repair orders, vendor invoices, credit memos, labor timecards, and sublet receipts to surface the money the shop is quietly leaving on the table — unreturned cores, missed vendor credits, duplicate charges, unbilled labor, missing shop supply fees, under-marked-up sublet, warranty-rate shortfalls, and diag-fee write-offs. The output is a prioritized recovery worksheet with dollar estimates, the RO or invoice to work from, the recommended action, and the owner/manager decision it needs.

Turn a shop's open-bench state (who you need, why, by when) into a complete multi-touchpoint recruiting package for one candidate or one role: a cold first-touch message, 2–3 follow-ups spaced over 10–14 days, a shop-culture job posting, an interview invitation and interview follow-up, and a working-interview / ride-along invitation. The output is built to work whether the candidate is actively applying (warm inbound) or a passive technician the shop is approaching directly (stealth outreach). The goal is to stop the "we'll hire when someone quits" cycle that leaves bays empty for 60+ days and replace it with a steady pipeline conversation.

Compile vehicle info, failure details, mileage, service history, and technician narrative into a complete, rejection-resistant warranty claim packet — formatted for manufacturer, extended warranty administrator (e.g., CarShield, Endurance, CNA, Zurich), or parts supplier warranty (e.g., NAPA, AutoZone Duralast, AC Delco). The output is structured so an office manager can paste it directly into the claim portal or email to the adjuster.

Turn rough notes into a polished, send-ready email tuned to the specific audience an auto-repair shop writes to every day — customers, insurance adjusters, warranty administrators, parts vendors, fleet accounts, manufacturer reps, landlords, regulators, or internal staff. Output includes a subject line, a properly structured body in the shop's voice, a clear next step, and a sender-notes block flagging follow-up timing and channel choice.

Transform raw meeting notes — voice-memo transcript, advisor handwriting, foreman bullet points — into a structured summary with clear decisions, action items with named owners and concrete deadlines, follow-ups, and parking-lot items. Designed for the meetings auto-repair shops actually run: morning huddles, weekly production planning, parts-ordering reviews, comeback / quality reviews, owner-manager 1-on-1s, fleet-account check-ins, and shop-wide all-hands.

Fast, single-pass review replies for the shop's routine review queue — 4-star and 5-star reviews and simple acknowledgments — so the owner, manager, or marketing coordinator can clear a weekly stack in one short session. For complex cases (1–3 star complaints, disputed facts, suspected fake reviews, platform-specific escalation rules) use the full-featured **Review Response Generator** in `customer-service/`.

Build a single monthly customer email newsletter — subject line, preview text, masthead, three to five sections, and a single primary CTA — that fits the shop's voice, the season, and the local customer base. The newsletter is the durable monthly touchpoint between visits: a seasonal-service reminder, a piece of customer education, a piece of shop news, an optional state-inspection or recall awareness note, and a single clear next step. Output includes the publish-ready HTML-ready Markdown body, a plain-text fallback for SMS / RCS where the shop sends in both formats, an alt-text suggestion for any embedded image, and a UTM-tagged primary CTA link for click attribution. One issue per call so the shop can run the skill once per month without producing a backlog.

Generate a single Google Business Profile post — text, photo brief, and a single clear call-to-action — tuned to the shop's locale, the current season, the services being marketed, and the brand voice. Output is a post body within Google's 1,500-character limit (300–700 characters is the practical sweet spot), a photo brief telling the shop owner what to shoot or pull from the gallery, an alt-text suggestion for accessibility, the post type (What's New / Offer / Event / Update), and a UTM-tagged booking link for click attribution. Two-to-four posts per month at biweekly cadence is the practical baseline; this skill produces one post per call so the shop can run the skill weekly without producing a backlog.

Generate a publish-ready 30–90 second short-form video script for Reels, TikTok, YouTube Shorts, or Facebook Reels — with a hook in the first 3 seconds, three to five visual beats with the on-camera dialogue and the b-roll cue, an optional caption, a single CTA, and a publish-ready caption text and hashtag set. The script is engineered for the auto repair shop's specific service-explainer use cases (coolant flush explainer, ADAS calibration explainer, parasitic draw explainer, brake-fluid moisture explainer, why-your-check-engine-light-isn't-always-urgent explainer, EV high-voltage safety explainer for general awareness, etc.) and respects the technician's actual time on camera (60 seconds of dialogue, max). Different from a full produced video — this is the script and shot list the shop owner or technician hands to whoever is filming, which is usually the same person on a phone in the bay between jobs.