ColdProof — claim-defense packet builder for small reefer carriers

1. One-liner

Turns a reefer download into a broker-proof rebuttal so small carriers stop eating spoilage chargebacks.

2. Trend signal — why now?

Three things collided in the last 18 months.

First, the money got worse for the people least able to defend themselves. Average refrigerated load values run $15,000–$85,000, and “one preventable excursion can wipe out a month of margin” (Source Logistics / industry guides, 2026). Big grocery and foodservice receivers apply chargebacks for rejected deliveries, and repeat rejections trigger contract reviews and supplier disqualification. The carriers absorbing this are overwhelmingly small: 91.5% of US carriers run 10 or fewer trucks and 23% of tractors are run by fleets with fewer than five (owner-operator statistics, 2026). These outfits have no claims department.

Second, the paper trail just became mandatory. FSMA-era food-safety practice now treats paper logs as insufficient; “missing records themselves constitute a violation,” and whether a carrier is held liable “depends on whether temperature records show compliance, document an excursion, or contain gaps” (Tive / cold-chain guides, 2026). Every reefer already logs setpoint, supply-air, and return-air data — but reading it correctly to win a claim is a specialist skill, which is why educational pieces like “Claims 101 – Interpreting Reefer Downloads” exist at all (United World Transportation, 2025).

Third, the data is now machine-readable and the parsing is now cheap. Thermo King TracKing exports Excel/PDF/RTF; Carrier Transicold and DAS loggers print or export the same readings; Motive/Lynx telematics expose them via API (Thermo King, Motive reefer monitoring). Document-AI can read those exports and apply the same return-air-vs-supply-air logic a claims expert uses — for cents.

Provenance:

• Signal 1 (demand): Owner-operators report $23k claims surfacing 7 months post-delivery, brokers holding pay, “a claim can be catastrophic for solo owners” — multiple TruckersReport threads — reefer load claims, broker lied about reefer temperature — accessed 2026-06-09
• Signal 2 (feasibility): Reefer downloads export to Excel/PDF/RTF and via TK/Carrier API; interpretation is a documented expert skill AI can now replicate — Thermo King USB datalogger, Claims 101: interpreting reefer downloads — 2025/2026
• Signal 3 (economic/regulatory): FSMA treats paper logs as insufficient and gaps as violations; chargebacks “wipe out a month of margin,” repeat rejections cause disqualification — Tive, Source Logistics — 2026 Category: Workflow automation

3. The opportunity

When a reefer load gets rejected, the carrier is presumed guilty until they prove the cold chain held. The proof exists — it’s sitting in the reefer’s own download — but turning that raw log into a coherent, defensible rebuttal requires knowing which sensor matters (return-air for “ran too warm” claims, supply-air for “ran too cold”), correlating it against the BOL setpoint, spotting whether the door was opened at the receiver, and writing it up in language a broker or cargo insurer will accept. A big carrier has a claims analyst who does this. A three-truck reefer outfit has a guy driving who finds out about a $23,000 chargeback seven months later via a deduction from settlement.

The incumbents make this worse, not better. MyEZClaim / CarrierClaim (TranSolutions) — the world’s first freight-claim software, now owned by MercuryGate — is built for shippers, brokers, and 3PLs to FILE claims and for large claims departments to manage them (MyEZClaim, MercuryGate acquisition). In other words, the existing tooling arms the side attacking the small carrier. Telematics players (Carrier Lynx, Tive, Motive) sell live monitoring and prevention, not after-the-fact defense, and they assume you’ve bought their hardware. Nobody sells the small carrier a cheap, fast “here’s why this claim is wrong, here’s your packet” tool that works off whatever reefer they already run.

That’s the 10× gap: collapse a specialist’s half-day of log interpretation and rebuttal-writing into a two-minute upload, priced for a wallet that’s already bleeding.

4. Target market

• Primary customer: Owner-operators and small refrigerated fleets (1–20 trucks) in the US hauling produce, meat, dairy, frozen, and pharma. The buyer is the owner or the office manager/dispatcher who handles settlements and broker disputes.
• Why they buy: In their words — “Sounds like they had sketchy product to begin with and they wanted you to take the hit for it!” and “The broker misrepresented the temperature for the load. It was supposed to be Fresh, it is Frozen at -10.” (TruckersReport). They feel railroaded, they have the data to fight, and they don’t know how to assemble it before the 21-day resubmission/dispute clock runs out.
• Rough TAM reasoning: Of ~922k owner-operators and the broader small-carrier base, reefer is a large, distinct segment (non-containerized reefer held 85% of the US refrigerated trucking market in 2025). Conservatively, 80,000–150,000 small reefer operators in the US. A reefer carrier sees several disputable temperature/condition claims a year; even a low-single-digit penetration at a modest subscription is a real business.
• Why now for them: Freight rates are deflated, insurance and fuel costs are elevated, small fleets “absorbed the brunt of rate deflation” and reefer segment failures are rising (Oliver Wyman, Luna Logistics). A single won claim is the difference between a profitable and a losing month.

5. Product sketch (MVP)

• Upload anything: drag in the reefer download (TK TracKing Excel/PDF, Carrier export, DAS/TouchPrint printout photo), the BOL, the rate confirmation, and the rejection notice.
• Verdict in plain English: “Your return-air held at 34–36°F against a 34°F setpoint for the entire haul. The excursion shown in the receiver’s claim occurred after the door-open event at delivery. This claim is not supported.” — with the exact data points cited.
• Conflict catcher: flags the classic killer — rate-con says 34°F but BOL says 0°F — and tells the carrier which one governs and what it means for liability.
• Defense packet (PDF): a clean, dated rebuttal letter + annotated temperature chart + cited download excerpts, formatted to send to the broker, receiver, or cargo insurer.
• Deadline tracker: logs the rejection date and counts down the dispute/resubmission window so claims don’t lapse silently.
• Claim journal: keeps every dispute, outcome, and dollar amount — so the carrier can spot a broker or lane that keeps “losing” their loads.

6. AI angle — what’s load-bearing

Remove the AI and there is no product. The core work is (a) reading heterogeneous, semi-structured reefer exports and even photographed printouts — OCR + document understanding — and (b) applying expert interpretation logic (which sensor governs which allegation, correlating against setpoint and door events, distinguishing a transit excursion from a post-delivery one) and (c) generating a persuasive, citation-backed rebuttal in the register a claims adjuster respects. That’s three AI-native tasks — extraction, reasoning over the extracted data, and structured generation. A non-AI version is just a folder and a template, which is exactly the status quo that’s failing these carriers.

7. Localization angle (if any)

N/A — this is a US-first play. The wedge is FSMA documentation norms, US broker/receiver chargeback behavior, and US reefer hardware (Thermo King, Carrier Transicold). The same shape ports to EU cold chain later (CMR claims, EN 12830 data loggers), but forcing localization now would dilute a sharp US-only insight. Geography is deliberately a fresh one for this portfolio, not a forced India/SEA cut.

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

• Pricing: $39/mo solo (owner-operator, up to ~3 disputes/mo) and $129/mo fleet (up to 20 trucks, unlimited disputes + claim journal). Plus a $49 one-time “rush packet” for non-subscribers who just got hit and want one defense now — a no-brainer when the claim is $20k.
• ACV: ~$700–900 blended (mix of solo subs, fleet subs, and rush one-offs converting to subs).
• Rough math to $1M ARR: ~1,200 paying accounts at a $70 blended monthly = ~$1.0M. Against a base of 80k+ small reefer operators, that’s ~1.5% penetration.
• Rough math to $5M ARR: ~5,500 accounts, or fewer accounts plus an attach: a per-won-claim success fee (e.g. 5% of the chargeback you avoid, capped) and a paid telematics integration tier. Add cargo-insurance broker partnerships that bundle ColdProof as a loss-mitigation perk.
• Expansion path: rush one-offs → solo subs → fleet subs → add lanes/broker analytics → success-fee on large claims → white-label for cargo insurers and TPAs who want fewer frivolous payouts.

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

• Forum ambush (the sharpest channel): TruckersReport and InsideTransport have active, searchable threads of carriers describing exact claims they’re losing right now. Answer the live threads with a genuinely useful free download-interpretation, then offer the packet. These are pre-qualified buyers in acute pain. (The $23k-claim and “broker lied about the temperature” threads are real, indexed, and recurring.)
• Reefer-haul YouTube/TikTok creators: owner-operator influencers (reefer-specific channels) run sponsorships cheaply and their audience is the customer. One “watch me beat a $14k bogus claim with ColdProof” video is a demo and an ad at once.
• Cargo-insurance & factoring partners: insurers and freight factors both lose money on disputed reefer claims and both already have lists of small reefer carriers. Offer ColdProof as a co-branded loss-mitigation tool; they distribute, we service. Factors especially (they hold the settlements where chargebacks land).
• Cold-conference/association presence: state trucking associations and produce-haul groups; a 5-minute “how to read your reefer download” clinic sells itself.

If those four don’t deliver 100 paying accounts in a quarter, the demand thesis is wrong — and that’s the kill signal.

10. Build complexity — justification

Medium. The web app, auth, billing, PDF generation, and the LLM reasoning/generation are all off-the-shelf. The real work is the parsing layer — normalizing TK/Carrier/DAS exports plus photographed printouts into a clean temperature series — and encoding the interpretation rules correctly enough that the rebuttal holds up. That’s a few months of focused work with a domain advisor (an ex-claims analyst or reefer-savvy dispatcher), not a research project. Realistic v1 for a 1–2 person team with an advisor: 10–14 weeks.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	Helping a carrier present its own data in a dispute. Not legal advice — keep it as document assembly, not “we’ll represent you.”
Ethical — no harm / dark patterns	✅	Helps the under-resourced side present truthful evidence. We surface when the data doesn’t support them too.
Market exists (evidence above)	✅	Recurring forum claims, real dollar figures, incumbents serving the opposite side.
1–5 person team can build this	✅	1–2 builders + a domain advisor.
Launchable with <$50K / ₹40L	✅	Inference + standard SaaS infra; main cost is the advisor’s time.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	16/20	$15k–$85k loads, claims that wipe a month’s margin, “catastrophic for solo owners.” Hair-on-fire when it happens — but episodic, not daily, which caps it below 17.
Demand evidence	15	11/15	Multiple independent forum threads with real dollar amounts and incumbents serving the filer side. Strong, but no public search-volume or revenue proof of carriers paying specifically for a defense tool yet.
Build feasibility	15	11/15	Mostly off-the-shelf; the parsing/interpretation layer across hardware formats and photographed printouts is the gnarly part.
Distribution clarity	15	11/15	Named channels (specific forums, reefer creators, insurers/factors) with pre-qualified buyers. Conversion math is estimated, not yet tested.
Revenue mechanics	15	11/15	Pricing fits the wallet and the rush one-off is a clean entry; $1M needs ~1.5% penetration, which is reasonable but unproven.
Time to first revenue	10	8/10	The $49 rush packet can sell during the build via concierge delivery — revenue in weeks, not months.
Defensibility	10	6/10	Soft moat: accumulating interpretation accuracy, a library of won-claim language, and insurer/factor distribution lock-in. Copyable, but a focused head start + the boring parsing grind is a real barrier.
Total	100	74/100

13. Qualitative modifiers

Founder-fit tags

technical-heavy (document-AI/parsing across messy formats) · domain-expertise-required (reefer download interpretation and claims norms — get an advisor day one).

Key assumptions to validate (3–5)

1. Assumption: Small reefer carriers will pay $39–129/mo (or $49 once) to win disputes that hit only a few times a year. How to test: Sell 15 rush packets concierge-style (you interpret by hand) to carriers from live forum threads; measure pay-through and willingness to subscribe afterward.
2. Assumption: AI can interpret real-world reefer downloads accurately enough that the rebuttal actually holds with brokers/insurers. How to test: Run 30 historical disputed downloads past a claims advisor; compare ColdProof’s verdict and packet to the expert’s. Target ≥90% agreement.
3. Assumption: A meaningful share of rejected reefer claims are actually defensible (otherwise we have nothing to sell). How to test: In the same 30-download sample, count how many the data exonerates or materially weakens. Need ≥40%.
4. Assumption: Insurers/factors will distribute. How to test: Pitch 5 cargo insurers / freight factors on a co-branded loss-mitigation tool; get one to agree to a pilot intro list.

Risk flags

1. Episodic usage / churn: Claims aren’t monthly, so a pure subscription may churn. Mitigate with the rush one-off as the front door and the claim-journal/deadline-tracker as the everyday reason to stay.
2. “Is this legal advice?” line: Stay strictly on document assembly and data presentation; don’t drift into representing the carrier or you invite UPL and E&O exposure.
3. Format fragmentation / platform dependency: New reefer firmware and telematics export formats keep changing; the parsing layer is permanent maintenance. Telematics APIs (Motive/Carrier) could also gate access.
4. Two-sided perception: Brokers/receivers may resist packets generated by a “carrier defense” tool. Counter by making outputs factual and neutral-toned, citing raw data — the goal is credibility with adjusters, not aggression.

14. Structured verdict

Score:                  74/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       Technical founder (document-AI) + a reefer-claims domain advisor
Time to revenue:        4–8 weeks (concierge rush packets during build)
Capital to launch:      $8–20K (advisor time + inference + SaaS infra)
Top 3 assumptions to validate first:
  1. Carriers pay for episodic claim defense — sell 15 concierge rush packets from live forum threads
  2. AI interpretation matches an expert ≥90% on 30 historical downloads
  3. ≥40% of real rejected reefer claims are actually defensible from the data
Kill criteria:
  - Abandon if <5 of 15 concierge rush-packet offers convert to paid
  - Abandon if expert-agreement on the 30-download benchmark is <80%
  - Abandon if <30% of sampled rejected claims are defensible from the download

15. Next step — 1-week validation sprint

• Day 1–2: Pull 30 real disputed reefer downloads + their rejection notices (from forum posters willing to share, plus a friendly small fleet). Recruit one reefer-claims advisor for a paid afternoon.
• Day 3–4: Hand-interpret each with the advisor; draft the packet for each manually (no product yet). Record: how many are defensible, what the winning argument was, how long each took.
• Day 5: Offer 15 active forum posters a $49 concierge “we’ll build your defense packet today.” Go/no-go: ≥5 pay AND advisor-defensible rate ≥40% AND your manual packets match the advisor’s call ≥90% of the time. Anything less, the wedge isn’t real yet.

The result is falsifiable: real dollars from real carriers, or a measured “no.”