WonBack — declined-repair recapture engine for independent auto shops

1. One-liner

Re-pitches every declined repair with the shop’s own photos until the customer says yes — no premium DVI platform required.

2. Trend signal — why now?

Three things are all true at once in mid-2026:

• The leak is quantified and loud. Industry data now puts declined recommended repairs at 30–40% of all recommendations, and shops that do systematically revisit them see 15–30% ARO lifts. Bolt On Technology frames missed approvals as “one of the most dangerous and least visible revenue leaks in auto repair.” Per-shop numbers being thrown around: $750K–$2.2M/year in declined work nobody follows up on.
• Trust — not price — is the real reason for the “no”. AAA research: two out of three American drivers don’t trust repair shops. The trade press is explicit that customers decline because “they can’t see the problem, feel pressured when a mechanic lists problems, and can’t review it later by the time they get home” — not because they can’t afford it.
• The follow-up is now a software problem, not a labor problem. Cheap multimodal AI (2025–26) can read a photo of a paper worksheet or a management-system export, understand “left front brake pad 2mm, rotor scored,” and generate a plain-English, urgency-framed re-pitch with the tech’s own photo attached. That was a human-writing task 18 months ago.

Provenance:

• Signal 1 (Demand): 30–40% of recommended repairs declined; “missed approvals… largest hidden revenue leak”; $750K–$2.2M/yr per shop un-followed-up — Bolt On Technology / whyoptimize — https://blog.boltontechnology.com/missed-repair-approvals + https://www.voicedrop.ai/auto-service-drive-retention-declined-repairs/ — 2026-07-01
• Signal 2 (Feasibility): Multimodal AI makes voice/vision viable for B2B field workflows; DVI photos already drive 15–30% ARO lift when revisited — BetterCloud SaaS-2026 / whyoptimize — https://www.bettercloud.com/monitor/saas-industry/ — 2026-07-01
• Signal 3 (Economic): Auto repair software market $3.4B (2026) → $8.6B (2033), 14.2% CAGR; Tekmetric (11,000 shops), BizzyCar, VoiceDrop all shipping declined-service recovery — Ratchet+Wrench / gitnux — https://www.ratchetandwrench.com/site-placement/latest-news/article/55319724/ — 2026-07-01 Category: Workflow automation

3. The opportunity

DVI platforms (AutoVitals, Tekmetric, Bolt On, Mitchell1) already capture the recommendation and, at the top tier, bolt on declined-service follow-up. Two gaps they leave open:

1. They’re bundled and priced for the 10+ bay shop. AutoVitals is repeatedly flagged as “above-average pricing” and “too in-depth / time-consuming to learn.” A 2-bay independent that still writes recommendations on a paper worksheet or texts them manually gets zero recapture — the declined line just walks out the door.
2. The follow-up that ships is a dumb reminder. “You still need brakes” — a generic SMS. It doesn’t re-show the visual evidence, doesn’t escalate urgency as the safety item ages, doesn’t speak like a human. The trade press is clear the “no” is a trust problem, and a bare reminder doesn’t solve trust.

WonBack is a standalone recapture engine: bring your own estimate (photo of the paper sheet, a Tekmetric/Shopmonkey CSV, or a forwarded text), and it runs the personalized, photo-backed, urgency-escalating re-pitch campaign the big platforms only give their premium tier — priced for a shop that grosses $400 ARO, not $700.

4. Target market

• Primary customer: Owner/operator of an independent US general-repair or specialty (brakes/tires) shop, 1–5 bays, 1–4 technicians, running on paper, basic texting, or a light SMS from a management system that has no real declined-service recovery. ~$400–550 ARO.
• Why they buy: “I know I’m leaving money on the table on every ticket I don’t follow up, but I’m the owner, the advisor, and sometimes the tech — I don’t have time to call people back and I forget who declined what.” The trade press confirms the operational reality: “reaching customers who previously declined services was nearly impossible without a firm process.”
• Rough TAM reasoning: ~160,000 US auto repair shops; independents hold ~70% (~112,000). Even if only the ~40–50% too small/cheap for a premium DVI bundle are reachable, that’s 45,000–55,000 shops. At $99/mo that’s a $50–65M ceiling — comfortably a sub-$5M-ARR play with a single-digit-percent slice.
• Why now for them: Parts and labor inflation squeezed margins; recapturing already-diagnosed work is the cheapest revenue in the shop. And for the first time the follow-up doesn’t need a $250/bay/mo platform or a hired advisor to write it.

5. Product sketch (MVP)

• Estimate ingestion, three ways: snap a photo of the paper worksheet, upload a CSV export from Tekmetric/Shopmonkey/etc., or forward the estimate text — WonBack parses line items, prices, and urgency.
• Auto-drafted re-pitch, per declined line: plain-English SMS + email that re-attaches the tech’s original photo of that customer’s part and explains why it matters, in the customer’s terms not shop jargon.
• Urgency-escalating cadence: safety items (brakes, tires, steering) get a tighter 1–2 week window and firmer language; wear items get a softer 2–4 week nudge. Caps at two touches so it never feels like spam.
• Reply handling: customer texts back “how much again?” or “can you do it Saturday?” — an AI agent answers with the quoted price and offers booking slots, escalating to the owner only on exceptions.
• Recapture dashboard: one number the owner checks — ”$ recaptured this month” and 14-day recapture rate — plus the list of open declined jobs ranked by $ value.
• Compliance-clean logging: every outreach attempt timestamped and stored, so the shop has a defensible record (“we told them the brakes were unsafe on X date”).
• No rip-and-replace: works alongside whatever the shop already uses. Doesn’t ask them to switch management systems.

6. AI angle — what’s load-bearing

Two places AI does the actual work:

1. Reading messy input into structured recommendations. A photo of a handwritten worksheet or an inconsistent CSV → normalized {part, measurement, urgency, price}. Without vision/LLM parsing this needs manual data entry, which is exactly the labor the shop doesn’t have. Remove the AI and the product is a spreadsheet the owner won’t fill in.
2. Writing a re-pitch that rebuilds trust, per customer. The whole thesis is that the “no” is a trust/comprehension problem, not a price problem. Generating a message that translates “LF rotor scored, 2mm pad” into “here’s the photo of your brake — this is why it squeals, and it gets more expensive to fix once it damages the rotor” — personalized to each declined line — is the load-bearing act. A mail-merge template can’t escalate urgency intelligently or answer the reply. If you strip the AI, you’re back to the generic “you still need brakes” SMS that already fails to convert.

7. Localization angle (if any)

N/A — this is a US play. The declined-service revenue leak, the DVI-platform pricing gap, and the paper-worksheet long tail are US-market-specific (labor rates, ARO norms, SMS-first customer comms). A UK/AU version is plausible later but there’s no localization wedge — the wedge is the underserved small-shop tier within the US.

8. Business model — path to $1M–$5M ARR

• Pricing: $99/mo flat per shop for v1 (single location, unlimited declined-job follow-ups). Later tier at $199/mo adds the AI reply-and-book agent and multi-location.
• ACV: ~$1,200–1,500/yr.
• Rough math to $1M ARR: ~700 shops × $99/mo × 12 ≈ $830K, plus a slice on the $199 tier → $1M. Out of a 45,000+ reachable pool, that’s <2% penetration.
• Rough math to $5M ARR: ~3,000–3,500 shops, blended ~$140 ACV, with the reply-and-book upsell attaching on ~40%. Requires a repeatable channel (see §9) and sub-4% monthly churn — the honest risk sits here.
• Expansion path: flat → per-location for multi-shop owners; usage upsell on the AI booking agent; a “recaptured revenue” success fee experiment (1–2% of booked recaptured work) that aligns price with proven value and lifts ACV without raising the base.

9. Go-to-market wedge — first 100 customers

• Scrape + personalized proof. Pull independent shops from Google Maps / Yelp in 3–4 metros, filter to those with <20 reviews mentioning “small shop / family owned” (a proxy for no premium software). Cold email/text the owner a 60-second Loom showing their own Google-listed services turned into a sample recapture message. Target 3–5% reply on a 2,000-shop list = 60–100 conversations.
• Ride the trade forums where owners already vent. AutoShopOwner.com and the Ratchet+Wrench community are dense with owner-operators discussing exactly this leak. Show up with a free “declined-revenue calculator” (upload last month’s declined lines → $ you left on the table), which doubles as the top of funnel.
• Parts-counter and jobber partnerships. Regional parts distributors (the WorldPac/NAPA jobber tier) already have the relationship with every small shop and a reason to want them selling more work. A rev-share referral through 2–3 regional jobbers reaches hundreds of shops without cold outreach.
• “Bring your declined list” free audit. First campaign free on the shop’s real backlog of declined jobs — recapture is measurable in 30 days, so the trial is the proof. Convert on the dollar figure it produced.

10. Build complexity — justification

Low. Everything is off-the-shelf: multimodal LLM for parsing + drafting, a texting/email provider (Twilio/SendGrid), a booking calendar, a thin dashboard. No hardware, no management-system integrations required for v1 (BYO-estimate sidesteps the integration tax). The only real engineering is making the parse reliable across messy handwriting/CSVs and tuning the urgency cadence. A pair ships a credible v1 in 8–10 weeks.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	Business-to-consumer SMS is legal with opt-out; shop is the sender-of-record for its own customers. TCPA-compliant consent flow required.
Ethical — no harm / dark patterns	✅	Re-pitching already-diagnosed, tech-recommended work with real photos; two-touch cap prevents harassment. Not manufacturing fake urgency.
Market exists (evidence above)	✅	Quantified leak, funded incumbents, growing software market.
1–5 person team can build this	✅	Off-the-shelf stack, 8–10 week v1.
Launchable with <$50K / ₹40L	✅	Solo/pair build + messaging costs; well under $50K.

All five pass.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	15/20	Real, recurring, dollar-quantified pain — but it’s a “leak” not “hair on fire.” Shops tolerate it today; switching cost from doing-nothing is low but so is urgency.
Demand evidence	15	12/15	Multiple independent signals: quantified decline rates, ARO-lift data, funded incumbents, trade-press consensus. Docked because the direct demand for a standalone small-shop tool (vs. bundled) is inferred, not yet proven.
Build feasibility	15	12/15	Off-the-shelf, 8–10 weeks. Docked for parse reliability across messy paper/CSV and TCPA-compliant messaging plumbing.
Distribution clarity	15	9/15	Named lists (Maps/Yelp scrape, AutoShopOwner, jobbers) and a measurable free-audit hook — but SMB owner-operators are notoriously slow to adopt and hard to reach at scale.
Revenue mechanics	15	10/15	Pricing benchmarked, ACV realistic, <2% penetration to $1M. Docked because churn on price-sensitive micro-shops is a real unknown and the $5M path needs a repeatable channel.
Time to first revenue	10	6/10	Recapture is provable in 30 days, so a paid pilot in 4–8 weeks is realistic — but the free-audit-to-paid conversion adds a cycle.
Defensibility	10	4/10	Weakest axis. Tekmetric, Bolt On, BizzyCar, VoiceDrop are all in-motion here; the moat is only the small-shop/BYO-estimate focus + accumulated recapture-message performance data. First well-funded incumbent that ships a cheap standalone tier compresses this fast.
Total	100	64/100

13. Qualitative modifiers

Founder-fit tags

sales-heavy (the win is repeatable SMB distribution, not the tech) · technical-heavy (parse reliability + messaging compliance).

Key assumptions to validate (3–5)

1. Assumption: Small independents will pay $99/mo for standalone recapture when their management system arguably “already does texts.” How to test: run the free-audit on 20 real shops’ declined backlogs; measure how many convert to paid after seeing 30-day recaptured $.
2. Assumption: A photo-backed, urgency-escalating re-pitch converts materially better than a generic reminder. How to test: A/B the two message styles on 500 real declined jobs across pilot shops; compare 14-day recapture rate.
3. Assumption: Owners will actually feed the tool their estimates (the input friction is survivable). How to test: track what % of pilot shops upload ≥1 estimate/week without nagging over 30 days.
4. Assumption: A regional jobber will do a referral rev-share. How to test: pitch 3 regional WorldPac/NAPA jobbers; one signed pilot referral partnership = validated.

Risk flags

1. Competitive compression: the biggest risk. Tekmetric (11,000 shops) or Bolt On unbundling a cheap standalone recapture tier would erase the wedge. Market is actively consolidating.
2. Platform/messaging dependency: TCPA and carrier SMS filtering (A2P 10DLC) can throttle deliverability; a compliance misstep is existential for a texting product.
3. SMB adoption drag: owner-operators who “don’t have time to follow up” also don’t have time to onboard software. Input friction (feeding estimates) could kill activation even where the value is real.

14. Structured verdict

Score:                  64/100
Verdict:                VALIDATE
Confidence:             Medium
Best-fit builder:       Sales-led operator who can grind SMB distribution, paired with one engineer for parse + messaging
Time to revenue:        6–10 weeks to first paid pilot
Capital to launch:      ₹3–6 lakh ($4–7K) — build + messaging + outreach tooling
Top 3 assumptions to validate first:
  1. Free-audit-to-paid conversion — run on 20 real shops, need ≥25% convert on 30-day recaptured $
  2. Photo-backed re-pitch beats generic reminder — A/B 500 declined jobs, need a clear recapture-rate delta
  3. Input friction survivable — ≥60% of pilot shops upload weekly without nagging
Kill criteria:
  - Abandon if <25% of free-audit shops convert to paid after seeing recaptured revenue
  - Abandon if a major incumbent (Tekmetric/Bolt On) launches a cheap standalone recapture tier before your v1 ships
  - Abandon if pilot-shop weekly upload rate stays below 40% (input friction wins)

15. Next step — 1-week validation sprint

• Day 1–2: Pull 300 independent shops in one metro from Google Maps; hand-build 15 sample recapture messages from their public service lists. Line up 10 owner conversations via cold text + AutoShopOwner DMs.
• Day 3–4: Get 3–5 shops to hand over their last month’s actual declined-job list. Run each through a manual (concierge, no product yet) recapture using their photos where available; send the messages.
• Day 5: Measure. Falsifiable outcome: across the pilot shops’ declined backlog, did the photo-backed re-pitch book ≥8% of declined jobs within the week, and did ≥2 of 5 owners say “yes, I’d pay $99/mo for this”? If neither, it’s a PASS — the leak is real but this cut doesn’t beat the incumbents’ bundled version.