ShortPay — supplement engine for independent collision shops

1. One-liner

Reads the insurer’s estimate, flags every owed-but-unbilled operation, and drafts the OEM-cited supplement the adjuster can’t refuse.

2. Trend signal — why now?

Three things moved at once.

First, the estimating databases keep shifting labor out of the “included” bucket. In the 2024–2026 guide updates, Audatex/Solera newly excluded seam-sealer application on new parts, assembly of multi-piece parts, riveting setup, and any labor to research OEM service information — each one now a line a shop must manually add or eat. The ASA publishes a running “Not-Included Operations” chart because there are too many to memorize. The DEG (Database Enhancement Gateway) has logged ~19,000 inquiries over 15 years about missing or wrong labor values; 52% resulted in a correction. Translation: the system underpays by design, and keeping up is a moving target no busy estimator wins.

Second, states are legislating the shop’s side of the argument. Texas SB 1429 (effective Jan 1, 2026) mandates OEM parts/procedures for newer insured vehicles and adds disclosure requirements. Rhode Island S 925 makes it an unfair claims practice for an insurer to refuse to compensate a shop for documented OEM- or paint-manufacturer-required procedures. Vermont signed a labor-rate study into law. Montana ran OEM-coverage bills. The legal scaffolding to back a supplement with a citation is being built right now.

Third, the money is provably on the table and quantified. Autobody News (John Yoswick) ran the math: a 150-RO/month shop missing one 0.3-hour operation on every estimate loses $1,890/month — $22,680/year — with no added staff or equipment. A separate survey found ~39% of shops aren’t fully paid even for OEM-required battery reconnection, and 28% never research OEM procedures at all.

Provenance:

• Signal 1 (demand): Shops leave $22,680/yr on the table from a single missed 0.3-hr operation per estimate; estimators “afraid to add operations for fear they’d be rejected” — https://www.autobodynews.com/news/common-estimating-mistakes-could-be-costing-your-shop-money — accessed 2026-06-01
• Signal 2 (feasibility): Estimating systems expanding “not-included” exclusions (seam sealer, riveting, OEM data research) in 2024–2026 guide changes — https://www.autobodynews.com/news/collision-repair-estimating-understanding-information-system-guide-changes — accessed 2026-06-01
• Signal 3 (economic/regulatory): States enacting OEM-procedure + labor-rate laws (TX SB 1429 eff. 2026-01-01; RI S 925 unfair-claims-practice; VT labor-rate study) — https://www.autobodynews.com/news/states-enact-legislation-related-to-oem-procedures-labor-rate-studies — accessed 2026-06-01 Category: Workflow automation

3. The opportunity

The incumbent isn’t a competitor product — it’s CCC ONE / Audatex / Mitchell, the three estimating platforms shops are forced to use because that’s where the insurer’s estimate lands. Those tools are neutral-to-insurer-aligned: they price the estimate, and when a shop downloads an insurer estimate or takes an “Open Shop” assignment, the insurer’s rates and rules ride along — the shop has to re-apply its own labor rate on every supplement, even supplement three, four, five, and the insurer keeps flipping it back. CCC ONE is also expensive enough that small shops ration its modules.

What none of them do is fight on the shop’s behalf. They won’t tell you “you performed a destructive weld test that’s a not-included op worth 0.4 hr — here’s the OEM doc number that proves it.” That judgment lives in the head of the one great estimator the shop probably doesn’t have. ShortPay productizes that estimator: it reads the insurer’s estimate, compares the documented damage and repair plan against the OEM procedures and the not-included rules for that exact estimating system, and hands back a finished, citation-backed supplement. We’re not replacing CCC — we sit beside it and recover the dollars CCC’s neutrality leaves behind.

4. Target market

• Primary customer: Owner/estimator at an independent (single-location or 2–4 location) collision repair shop in the US doing 60–250 repair orders/month. The person who writes estimates and runs the business and has no time to chase OEM PDFs.
• Why they buy: In their words — they’re “afraid to add operations for fear they’d be rejected,” the tech “stops to make a list of operations that were missed,” and the job gets “put on hold for insurance approvals.” They know they’re short-paid; they don’t have the hours to prove it line by line.
• Rough TAM reasoning: ~36,000 collision shops in the US (Autobody News’ stated reach). The Big Five consolidators + accelerators hold ~1/3 of revenue; roughly two-thirds of the market is still independent/small — call it 20,000+ shops that are exactly our buyer. At even 1,500 shops × $3,600/yr that’s $5.4M ARR without leaving the independent segment.
• Why now for them: The not-included exclusions just expanded again, and for the first time several states have put the shop’s OEM/labor-rate argument into statute. The justification they need to cite is now legally weightier than it was 18 months ago.

5. Product sketch (MVP)

• Drop in the insurer’s estimate (PDF/EMS export) + the shop’s photos and VIN.
• ShortPay decodes the VIN, pulls the applicable OEM repair procedures, and lists owed-but-missing operations keyed to the damage and the parts being replaced.
• For each flagged op: a plain-English reason, the matching not-included rule for the shop’s estimating system (CCC/Audatex/Mitchell), and a citation (OEM doc number, P-page, ASA not-included chart, or relevant state statute where applicable).
• One-click supplement draft — line items with labor units and justification language formatted to paste back into the estimate / email to the adjuster.
• A running recovery ledger: dollars flagged, dollars approved, approval rate by insurer — so the shop sees ROI per claim.
• Insurer-pattern notes (“Carrier X approves seam-sealer ops 80% of the time when cited to the P-page; pushes back on feather-prime-block”).

6. AI angle — what’s load-bearing

Remove the AI and there is no product — it’d just be the ASA PDF chart a shop already ignores. The model does three jobs a junior estimator can’t: (1) read an unstructured insurer estimate and the photos and infer what repair operations the damage actually requires; (2) reconcile that against the OEM procedure set and the current not-included rules for the specific estimating system — a reasoning task over messy, frequently-changing documents; (3) draft defensible justification language that cites the right authority for that op and that carrier. The whole value is collapsing a 45-minute blueprinting-and-research task into a two-minute review. That is squarely an LLM-over-documents job that became reliable in the last 12 months.

7. Localization angle (if any)

N/A — this is a US-only play, and deliberately so. The wedge is the US-specific machinery: CCC/Audatex/Mitchell databases, the DEG, the ASA not-included charts, and state-by-state OEM/labor-rate statutes. The same shop pain exists in the UK/EU/Australia, but the estimating systems, the legal hooks, and the “not-included” taxonomy are entirely different — that’s a later market, not a localization toggle.

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

• Pricing: $299/mo flat per shop (unlimited estimates) for the core; $499/mo “pro” tier adding the insurer-pattern analytics and multi-location ledger. No per-claim metering — shops hate variable bills and it kills usage.
• ACV: ~$3,600–$6,000.
• Rough math to $1M ARR: 280 shops × $300/mo × 12 = $1.01M. That’s <1.5% of the independent segment.
• Rough math to $5M ARR: ~1,200 shops at a blended $350/mo, or 800 shops if pro-tier mix climbs and we add a per-location upsell for small MSOs. Entirely inside the independent two-thirds of the market.
• Expansion path: seats/locations as small MSOs grow; an “audit-back” report product (proves recovered dollars for the shop’s books); eventually a paid carrier-pattern data feed. ACV grows from single-shop SaaS toward $1k+/mo for 3–5 location operators.

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

• DEG / SCRS / ASA orbit: the people filing those 19,000 DEG inquiries are our exact buyers and they congregate in known places (Database Enhancement Gateway, SCRS membership, state association meetings). Sponsor/post a free “what your last 10 estimates left on the table” teardown — run their real estimates, show the missed dollars, convert.
• Cold, value-first outreach: scrape state shop-licensing directories + Google Maps for independent shops; send each a 90-second Loom that runs ShortPay on a blurred sample of their own public review/photos showing 2–3 likely missed ops and an estimated annual leak. “Here’s $18k/yr you’re probably losing” is a reply-getting subject line; expect 4–6% reply at a few thousand sends.
• Estimator influencers + trade press: Mike Anderson / Collision Advice, FenderBender, Autobody News, and YouTube blueprinting channels already preach “stop leaving money on the table.” A tool that does it is a natural segment/sponsorship — co-branded “missed-ops audit.”
• Jobber/management-software resellers: integrate as an add-on with shop management systems that aren’t CCC; they have the install base and a reason to sell against CCC’s neutrality.

10. Build complexity — justification

Medium. Off-the-shelf: LLM document reasoning, VIN decode, PDF/EMS estimate parsing, web stack. The custom work is the operation→OEM-procedure→not-included-rule mapping and the citation library, kept current as the estimating guides change quarterly — that’s a real data-curation job and the source of the moat, but it’s curation, not research. A 2–3 person team (one strong builder + one collision-domain estimator on retainer) ships a credible v1 covering the top 10 makes in 4–5 months.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	Shops are entitled to bill for performed, documented operations; we draft justification, we don’t fabricate work.
Ethical — no harm / dark patterns	✅	Only surfaces ops supported by OEM docs + actual damage; explicitly flags “verify performed” so it never invents billing.
Market exists (evidence above)	✅	$22,680/yr quantified leak, 39% under-paid on OEM ops, expanding exclusions, new state laws.
1–5 person team can build this	✅	2–3 people, 4–5 months.
Launchable with <$50K / ₹40L	✅	Inference + data curation + a domain advisor; well under $50K to first revenue.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	16/20	Quantified recurring revenue leak felt on every job; not hair-on-fire (shops survive without it) but money + fear of rejection are real and weekly.
Demand evidence	15	12/15	Multiple independent signals: sourced dollar math, survey of under-payment, DEG inquiry volume, expanding exclusions. Direct “I’d pay $300/mo” still to validate.
Build feasibility	15	11/15	Standard stack + real, ongoing data-curation burden across estimating systems and makes.
Distribution clarity	15	12/15	Named channels (DEG/SCRS/ASA, trade press, scraped directories) with a killer “here’s your leak” hook; conversion math still a guess.
Revenue mechanics	15	12/15	Flat $299–499/mo benchmarks below CCC; ROI is self-evident if even one supplement/month clears. Retention risk if approval rate disappoints.
Time to first revenue	10	8/10	Shops buy fast on monthly SaaS; pilot-to-paid in weeks once the audit lands.
Defensibility	10	5/10	Execution + curated citation library + accumulating carrier-pattern data = soft moat. CCC could bolt this on; head start and shop-side alignment are the edge.
Total	100	76/100

13. Qualitative modifiers

Founder-fit tags

technical-heavy · domain-expertise-required — needs a builder comfortable with LLM-over-documents AND a real collision estimator who knows where the bodies (and the not-included ops) are buried.

Key assumptions to validate (3–5)

1. Assumption: Independent shops will pay $299/mo flat for recovered supplement dollars. How to test: run free audits on 30 real estimates from 30 shops, then ask for a paid pilot; track conversion.
2. Assumption: Citation-backed supplements actually lift the approval rate vs. the shop’s current ask. How to test: A/B 100 supplements (ShortPay-cited vs. shop’s normal) across a few pilot shops; measure approved dollars.
3. Assumption: The operation→OEM→not-included mapping can be kept current for the top makes without a research team. How to test: curate top-10 makes, then measure curation hours per guide-update cycle for one quarter.
4. Assumption: We can parse CCC/Audatex/Mitchell estimate exports reliably enough. How to test: ingest 200 real estimate exports, measure clean-parse rate.

Risk flags

1. Platform dependency: CCC/Audatex/Mitchell own the estimate format and could change exports or bolt on a competing feature; we depend on their files and rules.
2. Carrier pushback / relationship risk: if insurers start auto-rejecting “AI-generated” supplements, perceived value drops — value must come from correctness, not volume.
3. Curation drift: the not-included rules and OEM procedures change quarterly; stale citations erode trust fast. The moat and the maintenance burden are the same thing.

14. Structured verdict

Score:                  76/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       Technical founder (LLM-over-documents) + collision-estimator domain advisor
Time to revenue:        6–10 weeks from MVP (monthly SaaS, fast shop buying)
Capital to launch:      $25–40K ($ for inference, curation, and a domain advisor)
Top 3 assumptions to validate first:
  1. Shops pay $299/mo flat — free-audit-to-paid-pilot conversion on 30 shops
  2. Cited supplements lift approved dollars — A/B 100 supplements across pilots
  3. Top-10-make citation library is maintainable solo — track curation hours/quarter
Kill criteria:
  - Abandon if <15% of 30 audited shops convert to a paid pilot
  - Abandon if cited supplements don't measurably beat the shop's own ask on approved dollars
  - Abandon if a major estimating platform ships an equivalent shop-side feature before our v1

15. Next step — 1-week validation sprint

• Day 1–2: Collect 15–20 real recent insurer estimates (PDF/EMS) from 8–10 independent shops via DEG/SCRS forums and direct asks. Build a throwaway pipeline: parse estimate → flag missed ops against a hand-built not-included + OEM map for two popular makes.
• Day 3–4: Run the flagger on each estimate; with the domain advisor, hand-verify the flagged ops and total the recoverable dollars per shop. Send each shop their own leak number and a sample cited supplement line.
• Day 5: Ask each shop two things: “Is this number believable?” and “Would you pay $299/mo to get it automatically?” Go if ≥4 of 10 shops both confirm the dollars as real and verbally commit to a paid pilot. Falsifiable: a number and a yes/no, not a vibe.