PawProof — insurance claim-readiness concierge for vet clinics

1. One-liner

Reads a pet’s full record, drafts the history and pre-existing justification, and hands vets a first-pass-payable insurance packet.

2. Trend signal — why now?

Pet insurance is the fastest-growing line in personal P&C, and the claim — not the policy — is where the friction lives. The UK market was USD 2.04B in 2024 and is forecast to hit USD 5.0B by 2030 at a 16.2% CAGR (Grand View). US premiums cleared USD 4.8B in 2025 with claims volume up ~18% YoY (NAPHIA). More policies means more claims, and more claims land on the clinic’s desk because UK insurers settle directly with over 90% of practices (Petplan) — the vet is now an unpaid claims clerk.

The burden is now explicit and getting monetised. The Pet & Equine Insurance Association’s CEO Sharon Brown publicly noted that clinics are increasingly charging admin fees specifically for “providing claim history, notes or verification admin, supplying copies of clinical records and out-of-hours handling.” When a service becomes fee-worthy, it’s a budget line — and a budget line is something a SaaS tool can capture.

The denials are a documentation problem, not a coverage problem. Insurers reject claims for “insufficient medical records” and “pre-existing condition” far more than for genuine non-coverage (NAPHIA, Money.com). Pre-existing determination requires reading the entire longitudinal record and arguing chronology — exactly the long-context reading-and-summarising task that became cheap in the last 12 months. The carrier side is already automating its half: MGAs using AI invoice-verification report 55–65% straight-through processing vs 15–25% manual (Insurnest). The clinic side has no equivalent. That asymmetry is the opening.

Provenance:

• Signal 1 (Demand): PEIA CEO confirms clinics charging admin fees for claim history/records/verification; denials driven by insufficient records & pre-existing — https://www.insurancetimes.co.uk/analysis/are-pet-insurance-claims-keeping-pace-with-customer-expectations/1457580.article — 2026-06
• Signal 2 (Feasibility): Cheap long-context LLMs can read full PIMS records and draft pre-existing/medical-necessity justification; vendor-agnostic PIMS exports now common (SignalPET 150+ integrations) — https://www.signalpet.com/articles/the-top-10-veterinary-software-tools/ — 2026
• Signal 3 (Economic): UK pet insurance USD 2.04B→5.0B by 2030 (16.2% CAGR); US >USD 4.8B premiums, claims +18% YoY — https://www.grandviewresearch.com/industry-analysis/uk-pet-insurance-market-report — 2026 Category: Underserved niche

3. The opportunity

The claims workflow is being automated from both ends except the middle. Carriers automate adjudication (Five Sigma, Insurnest). PIMS vendors built carrier-locked direct-pay rails (ezyVet↔Trupanion, Petplan direct settle). But both of those only work for the handful of insurers with a clinic integration. There are 50+ pet insurers in the UK and US; the long tail has no integration, and the clinic still does the work by hand: pull the record, decide what’s relevant, write the history summary, justify that the condition isn’t pre-existing, attach the itemised invoice, email or fax it.

That manual middle is what bounces. A claim that gets kicked back for “we need more history” costs the clinic two more touches and the client three more weeks of waiting — and the client blames the clinic, not the insurer. PawProof owns that middle: carrier-agnostic, denial-focused, working off the clinic’s own record regardless of which of the 50 insurers the client picked. The incumbents win the integrated insurers; we win everyone else — which is most claims.

The 10× isn’t “submit faster.” It’s “submit once.” First-pass payable is the entire value proposition.

4. Target market

• Primary customer: Practice manager / lead vet nurse at an independent or small-group (1–6 vet) veterinary clinic in the UK, then US/AU/Canada. The person who actually wrestles the insurance paperwork, not the clinical vet.
• Why they buy: “We spend hours a week chasing records and rewriting histories for insurers, claims bounce, and the client rings us angry when their reimbursement is late.” The clinic eats the time and the reputational hit for a process it doesn’t even profit from.
• Rough TAM reasoning: ~5,000 veterinary practices in the UK, ~30,000 in the US, ~30% of UK pets and a fast-growing US share insured. Even at a UK-first beachhead of 5,000 clinics × £150/mo, that’s a £9M ARR ceiling in one country before US/AU. The serviceable wedge — independent and small-group clinics that handle insurance in-house — is several thousand clinics paying real money monthly.
• Why now for them: Direct-settle culture made the clinic the default claims processor; rising claim volume turned a chore into a daily grind; clinics started charging admin fees, proving the work has explicit monetary value they’d happily offload for less than they’re spending.

5. Product sketch (MVP)

• Drop in (or auto-pull via export) a pet’s clinical record; PawProof reads the full longitudinal history, not just the visit being claimed.
• Generates a clean, insurer-ready claim history summary — the chronological narrative carriers demand, written to their evidentiary standard.
• Pre-existing-condition check: flags whether the presenting condition has prior mentions, builds the timeline, and drafts the argument for why it is or isn’t pre-existing — the #1 denial reason, handled head-on.
• Attaches and labels the itemised invoice and diagnostics in the format the target insurer expects (insurer profile library: required fields, common rejection triggers).
• First-pass risk score: before submission, flags the specific gaps an adjuster will bounce (“no recheck note for the chronic ear condition”) so the clinic fixes it once.
• One-click export packet (PDF + structured data) ready to email, fax, or paste into the carrier portal — carrier-agnostic by design.
• Claim tracker: which claims sent, which paid, which bounced and why — feeding the gap-detector to get smarter per insurer.

6. AI angle — what’s load-bearing

Remove the AI and there is no product — it’s a fax cover sheet. The load-bearing work is reading an unstructured, multi-year clinical record (free-text SOAP notes, abbreviations, inconsistent terminology) and producing two things a human currently spends 20–40 minutes on per complex claim: a defensible chronological history, and a pre-existing-condition determination with the supporting argument. That’s longitudinal reasoning over messy medical text — squarely an LLM job, and one that only became cheap enough to run per-claim in the last year. The insurer-specific gap-detection (“this carrier rejects without a follow-up note”) is a learned pattern layer on top. No model = no value.

7. Localization angle

UK-first is the wedge, not a footnote. The UK has the highest insured-pet claim frequency, an established direct-settle norm (insurer pays the clinic), and clinics already charging insurance admin fees — so the buyer is primed and the workflow is daily. Insurer profiles are local: Petplan, ManyPets, Bought By Many, Agria, Animal Friends each have their own required fields and rejection quirks. A generic global tool can’t encode that; a UK-tuned one wins, then ports the same insurer-profile machinery to the US (Trupanion, Lemonade, Healthy Paws, Nationwide), Australia, and Canada one market at a time. £/$ micro-pricing (£99–199/mo) fits a small clinic’s wallet where enterprise claims software does not.

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

• Pricing: £149/mo per clinic flat (unlimited claims) for small clinics; £299/mo for multi-site groups. Optional per-claim overage above a fair-use cap for high-volume practices.
• ACV: ~£2,200/year blended.
• Rough math to $1M ARR: ~450 clinics × £149/mo × 12 ≈ £800K ≈ USD 1M. ~9% of UK independent clinics.
• Rough math to $5M ARR: ~1,900 clinics across UK + US + AU, or fewer with group-account upsell and per-claim expansion at high-volume hospitals. Requires winning a second geography and one PIMS-marketplace listing.
• Expansion path: seats → multi-site group plans → per-claim overage → insurer-profile premium tiers → eventually charge for the “first-pass risk score” as a standalone audit on the clinic’s claim backlog.

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

• Veterinary nurse / practice-manager Facebook groups and forums (e.g. large UK vet-nurse and practice-manager groups, BVNA-adjacent communities). These are where the insurance-paperwork venting already happens. Post a 90-second before/after of a real bounced claim getting cleaned up; offer 20 free claim-preps to the first clinics that DM. Target 5% of an engaged 10K-member group.
• Cold-walk the local cluster: scrape the RCVS/Find-a-Vet directory (every UK clinic is listed), email the practice manager a personalised loom showing their type of claim being prepped. ~3,000 independent clinics; 3–5% reply on a specific-pain pitch is 90–150 conversations.
• Ride a denial complaint: monitor pet-owner forums and Trustpilot for “my vet had to resubmit / claim took weeks” posts, then approach the named clinic with “your clients are blaming you for the insurer’s bounce — here’s the fix.”
• PIMS marketplace listing (ezyVet/Provet/Vet Hub app stores) once v1 is proven — distribution to clinics already shopping for add-ons.
• One reference KPI to win on: “X% fewer resubmissions” from the first 10 design-partner clinics becomes the entire sales pitch.

10. Build complexity — justification

Medium. The reading-and-drafting core is off-the-shelf long-context LLM work over exported records — no custom model, no PIMS write-back needed for v1 (import + export packet is enough). The genuine work is the insurer-profile library (required fields, rejection triggers per carrier) and the clinical-text robustness needed to not hallucinate a pre-existing condition either way — that’s the part that earns trust and the part that takes engineering discipline and a vet advisor. Realistic v1 for a small team: 3–4 months, with the first month spent shadowing two design-partner clinics’ actual claim flow.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	Clinic processes its own client’s records with consent; no new regulated activity. Data-protection (UK GDPR) handled as standard processor.
Ethical — no harm / dark patterns	✅	Helps legitimate claims get paid faster; must avoid gaming pre-existing determinations — product flags chronology honestly, doesn’t fabricate.
Market exists (evidence above)	✅	Clinics already charging admin fees for this exact work; booming claim volume.
1–5 person team can build this	✅	LLM-over-exports + insurer-profile library; 3–4 months.
Launchable with <$50K / ₹40L	✅	No capex; inference + dev only.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	15/20	Real, weekly, money-and-reputation pain — but it’s the clinic’s secondary headache, not hair-on-fire like a frozen export licence.
Demand evidence	15	12/15	Strong: admin fees being charged, booming claims, named industry-body commentary. Light on verbatim clinic quotes — that’s the validation gap.
Build feasibility	15	12/15	Core is off-the-shelf; insurer-profile + anti-hallucination rigor is the real work.
Distribution clarity	15	12/15	Named directories, named forums, clear conversion math; PIMS marketplace as second channel.
Revenue mechanics	15	11/15	Flat £149/mo fits wallets and math closes at <500 clinics — but flat unlimited pricing caps expansion unless overage lands.
Time to first revenue	10	8/10	Clinics buy add-ons fast at this price; design-partner pilots can convert in 6–8 weeks.
Defensibility	10	5/10	Insurer-profile library + per-clinic claim-outcome data compound, but a PIMS vendor or a funded scribe could bolt this on. Execution + niche focus moat only.
Total	100	75/100

13. Qualitative modifiers

Founder-fit tags

technical-heavy · domain-expertise-required — needs solid LLM/eval engineering and a vet or ex-claims-handler advisor to encode insurer quirks and keep pre-existing calls honest.

Key assumptions to validate (3–5)

1. Assumption: Clinics spend enough staff time on insurance claim prep that £149/mo is an obvious saving. How to test: Time-and-motion interviews with 20 UK practice managers — get actual minutes-per-claim and claims-per-week.
2. Assumption: First-pass payable rate measurably improves with our packet. How to test: Run 50 real claims for 5 design-partner clinics through PawProof vs their normal process; compare resubmission rate.
3. Assumption: Carrier-locked direct integrations haven’t already absorbed the long-tail volume. How to test: Survey design partners on what % of their claims go to insurers with NO integration (the addressable slice).
4. Assumption: Pre-existing determination can be done reliably enough to be trusted, not just plausibly. How to test: Vet-graded eval set of 100 records with known outcomes; measure false pre-existing calls in both directions.

Risk flags

1. Platform/incumbent encroachment: A PIMS vendor (ezyVet, Provet) or a funded scribe (CoVet, VetRec) could add carrier-agnostic claim prep as a feature. Counter: move fast, own the insurer-profile depth and claim-outcome data they don’t have.
2. Market-squeeze timing: Carrier-side auto-adjudication and direct-pay integrations are both expanding; the addressable “manual middle” could shrink over 3–5 years. Counter: lead with the long-tail-insurer + pre-existing-denial slice that integrations don’t touch.
3. Trust/liability: A wrong pre-existing call could cost a client a claim or a clinic its credibility. Counter: position as a drafting assistant with human sign-off, never auto-submit; conservative flagging.

14. Structured verdict

Score:                  75/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       Technical founder (LLM/eval) + vet or ex-pet-claims-handler advisor
Time to revenue:        6–10 weeks to first paid pilot
Capital to launch:      £8–15K ($10–20K) — inference + dev only
Top 3 assumptions to validate first:
  1. Minutes-per-claim and claims-per-week — 20 practice-manager interviews
  2. Resubmission-rate improvement — 50 real claims, 5 design partners, A/B vs normal flow
  3. Addressable share — % of a clinic's claims going to insurers with no integration
Kill criteria:
  - Abandon if design-partner clinics spend <15 min/claim on prep (pain too small to pay for)
  - Abandon if PawProof packets don't cut resubmission rate by ≥30% in the 50-claim test
  - Abandon if >70% of a typical clinic's claim volume already flows through integrated direct-pay carriers (no addressable middle)

15. Next step — 1-week validation sprint

• Day 1–2: Recruit and interview 15–20 UK practice managers from vet-nurse Facebook groups. Capture minutes-per-claim, claims-per-week, resubmission frequency, and what % of claims go to non-integrated insurers.
• Day 3–4: With 2 willing clinics, take 10 real anonymised records and hand-run the PawProof flow (prompted, no UI) — produce the history summary + pre-existing determination. Show the managers; ask “would you pay £149/mo for this, and would it have prevented your last bounce?”
• Day 5: Go/no-go. Go if ≥10 of 20 managers report ≥15 min/claim AND ≥3 of the 2-clinic test packets would have prevented a real resubmission. Falsifiable: a hard count of prevented bounces, not “they liked it.”