PlanB

1. One-liner

Turns a declined vet estimate into a clinically-defensible cheaper plan before the client leaves the exam room.

2. Trend signal — why now?

There is a measurable, well-sourced gap between what veterinarians think they offer and what pet owners report receiving. That gap is the product.

Gallup surveyed 933 practicing U.S. veterinarians (fieldwork 9 Sep – 5 Oct 2025, published 20 Jan 2026, ±3.3pp). The findings:

• 94% say client finances often (38%) or sometimes (56%) prevent them from providing recommended treatment.
• 81% say they always (37%) or often (44%) recommend an alternative treatment plan when cost causes a decline.
• But in the companion pet-owner study, 73% of pet owners who declined care due to cost said they were not offered a lower-priced option.
• 98% are concerned the patient’s condition worsens or becomes chronic as a result.

Vets believe they offer a Plan B four times out of five. Owners say they heard one roughly one time in four. Somebody’s wrong, and the peer-reviewed literature says it’s the vets.

Dolan & Slater surveyed 1,160 veterinarians (Animals, May 2024). When financial limits were known, 91% offered a spectrum of care. When financial limits were not known — i.e. the ordinary case, before anyone says the word “money” — that collapses to 61%. Only about half reported offering simpler diagnostic options first. The single significant predictor in both statistical models was veterinarian comfort, not knowledge.

A separate U.S. survey found 88% of veterinarians are familiar with “spectrum of care” but only 45% say they always offer a full range of options. The barrier cited repeatedly is time: these conversations take minutes nobody has in a packed appointment block.

Meanwhile the economics turned hostile. Per iVet360 reporting covered by Today’s Veterinary Business, client visits dropped 4% in 2024 while average transaction charge rose 6.1%. Practices are covering falling volume with higher prices — which is exactly the condition under which a sticker-shocked owner walks out. Industry accounting estimates practices miss 5–10% of all charges, worth up to $150K/yr at a $1.5M practice.

The build got cheap in the last 18 months. Veterinary AI scribes went from novelty to installed base: Scribenote raised $8.2M seed led by a16z and serves 6,000+ vets; VetRec is YC-backed with 10+ PIMS integrations at $99/vet/mo. Search volume for “veterinary AI” grew over 1,680% YoY between 2024 and 2025 to ~335K monthly searches. The integration rails and the buying appetite both exist now and did not in 2023.

Provenance:

• Signal 1 (demand): Gallup — 94% of vets say cost blocks treatment; 81% claim to offer alternatives, but 73% of owners who declined say they were never offered a cheaper option — https://news.gallup.com/poll/700115/veterinarians-say-cost-main-driver-declined-care.aspx — 20 Jan 2026
• Signal 2 (feasibility): Veterinary AI scribes hit mainstream with mature PIMS integration rails — Scribenote $8.2M seed led by a16z, 6,000+ vets; VetRec YC-backed with 10+ PIMS integrations — https://www.scribenote.com/blog/scribenote-pioneers-ai-scribe-to-combat-burnout-among-veterinarians-as-it-secures-8-2m-seed-round-led-by-a16z — 2025/2026
• Signal 3 (economic): Client visits fell 4% in 2024 while transaction charges rose 6.1%; practices miss 5–10% of all charges (~$150K/yr at a $1.5M practice) — https://todaysveterinarybusiness.com/ivet360-report-042625/ — 25 Apr 2025
• Signal 4 (clinical validation): Dolan & Slater, n=1,160 — spectrum-of-care offering drops from 91% to 61% when client financial limits are unknown; vet comfort is the only significant predictor — https://pmc.ncbi.nlm.nih.gov/articles/PMC11117253/ — May 2024 Category: Tech-unlock

3. The opportunity

The declined estimate is the single worst moment in a veterinary practice’s day. The pet doesn’t get treated, the vet feels moral injury, the client feels judged, and the clinic books zero revenue on a full appointment slot. Then everyone pretends it didn’t happen.

The incumbent isn’t a competitor — it’s the blank space after “I’m sorry, I can’t afford that.” Today the alternative plan gets improvised, if at all, by a tired associate at 6pm who is not confident about which diagnostics are safely droppable. Dolan & Slater identified precisely this: it’s a comfort problem, not a knowledge problem. Vets don’t downgrade a plan because they’re afraid of practicing below standard of care and of the liability that implies.

The existing tools miss it from both sides. Instinct Standards publishes excellent evidence-based monographs with a spectrum-of-care lens — but it’s a reference library you consult, not something that reads this estimate for this patient and produces a costed Plan B in fifteen seconds. PIMS estimate builders (AVImark, Cornerstone, ezyVet) let you hand-build line items; they have no clinical opinion about what to remove. AI scribes (Scribenote, VetRec, HappyDoc) capture what was said after the fact and stop at the SOAP note. Nobody sits at the decision point.

The 10× isn’t “cheaper software.” It’s converting a walk-out into a $180 transaction while producing the medical-record documentation that makes the vet legally and psychologically safe doing it.

Important design constraint, learned from disconfirming research: Today’s Veterinary Business explicitly warns against naive tiering — “I’m not a fan of that approach because pet owners might think they should pick the services they want and that all the recommendations are somewhat equal.” A dumb good/better/best menu is a known anti-pattern in this profession. The product must present a clinically sequenced alternative — “here is what we defer and what we monitor instead, and here is the trigger that brings us back” — not a pricing tier list.

4. Target market

• Primary customer: Practice owner or hospital manager at an independent 1–4 DVM companion-animal general practice in the US, $800K–$3M annual revenue, running AVImark, Cornerstone, ezyVet, Shepherd, or Digitail.
• Why they buy: They watch estimates get declined weekly and book nothing. Gallup: 94% report cost blocking treatment. Their associates are burning out over it — 98% are concerned about patients deteriorating after a decline. A tool that recovers even two declined estimates a week at $150 each pays for itself six times over.
• Rough TAM reasoning: AVMA counts 30,000+ US veterinary practices; 82.6% are primarily general medicine and surgery. Corporates own 22% and are largely unreachable for a bootstrapper, leaving roughly 19,000–20,000 independent GP clinics as the serviceable base. At $199/mo, capturing 2.5% of that base is ~$1.15M ARR.
• Why now for them: Visits fell 4% in 2024 while they raised prices 6.1%. They cannot price their way out any further, and every declined estimate is now a bigger share of a shrinking book.

5. Product sketch (MVP)

• Plan B in one click — from an existing estimate in the PIMS, generate a clinically-sequenced lower-cost alternative with what’s deferred, what’s substituted, what’s monitored instead, and the trigger that escalates back to the full plan.
• Deferral rationale, written for the record — each omitted item carries a one-line clinical justification, so the medical record shows a reasoned decision rather than a discount.
• Practice-priced, not list-priced — pulls the clinic’s own fee schedule so the alternative reflects real local pricing, not a generic benchmark.
• Client-facing handout — plain-language summary the owner takes home explaining what’s being done, what’s being watched, and what symptoms mean “call us now.”
• Declined-estimate log — tracks what was declined, what Plan B was offered, and whether it was accepted. Gives the owner the first real price-elasticity data they’ve ever had.
• Recheck trigger queue — surfaces patients on a deferred plan whose recheck window is closing, so a downgraded plan doesn’t become an abandoned patient.
• Evidence citation per recommendation — links each substitution to published spectrum-of-care guidance, which is what makes the associate comfortable enough to actually use it.

6. AI angle — what’s load-bearing

Remove the AI and this is a spreadsheet nobody opens.

The load-bearing work is clinical reasoning under a cost constraint, executed in under thirty seconds: read a specific estimate for a specific signalment and presenting complaint, decide which diagnostics are safely deferrable versus which are load-bearing for the differential, propose substitutions (in-house chemistry instead of a send-out panel; empirical therapy with a defined recheck instead of imaging first), and write a defensible justification for each in medical-record language.

That’s a judgment task over a long tail of presentations. No rules engine covers it — the combinatorics of species × complaint × comorbidity × owner constraint are exactly why practices don’t have a playbook today. It’s also why the 2024 study found comfort was the binding constraint: the vet needs to see reasoning they’d endorse, with a citation, fast enough to use mid-appointment.

The honest risk: this is clinical decision support in a licensed profession. The product must be framed and built as a drafting aid the veterinarian reviews, edits, and approves — never as autonomous recommendation. That’s both an ethics requirement and a liability requirement, and it shapes the whole UX.

7. Localization angle

N/A — this is a US-first play. The wedge is the specific US structure of private-pay companion-animal medicine: no third-party payer for most visits, so the cost conversation happens directly between vet and owner in the exam room, and the practice eats 100% of a decline. In markets with high pet-insurance penetration (Sweden ~90%, UK ~25%) the decline dynamic is materially weaker. The UK is the obvious second market — its “contextualized care” movement is the same idea under a different name, and the CMA’s ongoing scrutiny of veterinary pricing transparency creates a parallel pull — but it’s a phase-two expansion, not the wedge.

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

• Pricing: $199/mo per practice for 1–2 DVMs; $349/mo for 3–5 DVMs. Practice-level, not per-seat — techs and front desk need access for the log and handouts, and per-seat pricing would suppress exactly the usage that creates value.
• Benchmark sanity check: PIMS themselves run $290–$549/mo (Shepherd $299 for one doctor, Neo $290, Cornerstone $549). AI scribes run $79–$150/vet/mo. A $199 practice-wide add-on sits comfortably inside an established software budget.
• ACV: ~$2,600 blended.
• Math to $1M ARR: 385 practices × $199/mo × 12 = $920K; with ~20% on the $349 tier, ~$1.02M. That’s 2% of the ~19,000 independent GP base.
• Math to $5M ARR: ~1,600 practices (8% of the independent base) plus expansion into corporate groups — a single mid-size consolidator with 200 hospitals at negotiated rates is worth $400–500K ARR alone. Realistically $5M requires winning 2–3 corporate accounts, which means the timeline stretches past 24 months.
• Expansion path: per-DVM tiers as practices grow → declined-estimate benchmarking across the customer base (what do comparable clinics recover?) → a payment-plan/financing referral cut, since Gallup shows only 41% of vets regularly offer payment options and that’s an obvious adjacent monetization.
• Margin note: inference is the main variable cost. A Plan B generation is a bounded reasoning task; at realistic volumes (30–80/mo per practice) this is single-digit dollars per account per month. Gross margin holds above 85%.

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

• The VMA state-conference circuit. There are ~50 state veterinary medical associations running annual CE conferences with exhibit halls. Booths at mid-size state VMAs run $1,200–$2,500. Practice owners attend specifically to shop tools. Work 6–8 conferences in year one; a working demo that takes a real estimate and produces a real Plan B in thirty seconds is a hallway-stopping demo. Target 15–25 trials per conference.
• Ride the spectrum-of-care movement, don’t invent a category. AVMA published a JAVMA spectrum-of-care supplement; AAVMC runs a Spectrum of Care Initiative across veterinary colleges; AAHA publishes on it actively. This is an organized professional movement actively looking for implementation tooling and finding only reference material. Sponsor/co-present CE webinars with the educators already evangelizing it — they have the audience and no product.
• Target new graduates and their employers. Dolan & Slater found vets with 20+ years’ experience were 3–4× less likely to offer spectrum of care than newer practitioners. New grads want this and are the least comfortable improvising it. MentorVet and similar early-career wellbeing programs are a direct channel to exactly the cohort with the pain and the least entrenched habits.
• The declined-estimate audit as a free wedge. Offer any practice a free one-time analysis: export last 90 days of declined estimates from the PIMS, return a report showing total declined dollars and what a Plan B would have plausibly recovered. This is a cheap, concrete, self-serving artifact — it quantifies their bleed in their own numbers and lands the trial. Run it on 200 clinics sourced from state licensing directories.
• PIMS marketplace listings. Covetrus Pulse alone lists 250+ third-party integrations, and Vetstoria/Otto/VitusVet have all proven the integration-partner path. Slower than direct sales but compounding, and it borrows the PIMS’s credibility.

10. Build complexity — justification

Medium. The generation engine is off-the-shelf frontier-model reasoning over a curated corpus of published spectrum-of-care guidance plus the practice’s own fee schedule — no model training required. The real work is PIMS integration: ezyVet has a documented REST API with OAuth 2.0, but AVImark is a client-server legacy application serving 11,000+ hospitals, and reaching it means either a partner agreement or a local agent. That integration surface is what pushes this from Low to Medium.

Realistic path: 10–12 weeks to a working v1 against ezyVet plus a manual paste-in fallback (paste the estimate, get a Plan B) that works for every practice regardless of PIMS. The paste-in path is what lets you sell before integrations are done. Add AVImark and Cornerstone over the following 3–4 months. A technical founder plus a veterinarian advisor (equity, not hire) is the right shape; the clinical corpus curation genuinely requires a DVM’s judgment.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	Clinical decision support presented to and approved by a licensed veterinarian. No autonomous recommendation, no VCPR created, no direct-to-consumer clinical advice. Must ship with clear professional-use framing.
Ethical — no harm / dark patterns	✅	Explicitly increases access to care for financially constrained owners — the alternative today is nontreatment or economic euthanasia. Guardrail required: the product must never propose a plan below reasonable standard of care, and must always surface the escalation trigger.
Market exists (evidence above)	✅	Gallup n=933, Dolan & Slater n=1,160, 30,000+ US practices, funded adjacent category (a16z, YC).
1–5 person team can build this	✅	Technical founder + DVM advisor; 10–12 weeks to sellable v1 via paste-in fallback.
Launchable with <$50K / ₹40L	✅	~$25–35K: conference booths ($10–15K), DVM advisor equity/stipend, inference, incorporation and professional-liability review.

All five pass.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	16/20	Felt weekly, costs real money, and carries genuine moral weight — 98% of vets worry about post-decline deterioration. Not 18+ because the practice survives without solving it; it’s a recovered-revenue play, not a compliance deadline. Nobody gets fined for skipping it.
Demand evidence	15	13/15	Unusually strong for a fresh idea: two independent surveys (n=933, n=1,160) quantifying the exact gap, a funded adjacent category, 1,680% YoY search growth in veterinary AI. Short of 15 because no one is yet paying for this specific product — demand for the outcome is proven, willingness-to-pay for the tool is inferred.
Build feasibility	15	11/15	Core reasoning is off-the-shelf. PIMS integration — especially legacy AVImark — is the tax. The paste-in fallback rescues the timeline and de-risks the launch.
Distribution clarity	15	12/15	Named channels with real math: ~50 state VMA conferences, an organized professional movement with no tooling, and the free declined-estimate audit as a concrete wedge. Not higher because veterinary sales cycles run slower than the outreach math suggests and owners are notoriously software-fatigued.
Revenue mechanics	15	12/15	Pricing benchmarked directly against PIMS ($290–549) and scribes ($79–150). 385 practices to $1M is credible. Docked for the $5M path genuinely requiring corporate deals, which is a different sales motion than the one that gets you to $1M.
Time to first revenue	10	7/10	Paste-in v1 can be pre-sold and generating revenue in 8–12 weeks. Slower than a pure self-serve tool because clinics buy after seeing it work on their own cases, and conference-driven pipelines are lumpy.
Defensibility	10	5/10	Honest score. The clinical corpus and accumulated declined-estimate/acceptance data compound into a real advantage by month 12 — knowing which Plan B variants actually get accepted is proprietary and valuable. But at month 3 it’s copyable, and any funded scribe company (Scribenote has a16z money and 6,000 vets) could bolt this on as a feature. Speed and a narrow focus are the only early moat.
Total	100	76/100

13. Qualitative modifiers

Founder-fit tags

technical-heavy · domain-expertise-required

This does not work without a practicing veterinarian deeply involved. The clinical corpus, the deferral logic, and the credibility to sell into VMA conferences all require a DVM. A pure software founder building this alone will produce something vets won’t trust — and trust is the product, per the comfort finding.

Key assumptions to validate (3–5)

1. Assumption: Practices will pay $199/mo for recovered revenue rather than treating declined estimates as an unavoidable cost of business. How to test: Run the free declined-estimate audit on 25 clinics; measure how many, on seeing their own quantified bleed, agree to a paid pilot. Need ≥6.
2. Assumption: Veterinarians will trust and actually use an AI-drafted alternative plan mid-appointment, given the comfort barrier is psychological rather than informational. How to test: Put a working prototype in front of 15 DVMs with real anonymized declined estimates. Measure: would they present this to a client as-is, with edits, or not at all? Need ≥60% “as-is or light edits.”
3. Assumption: Plan B acceptance is materially better than zero — that a downgraded plan converts a meaningful share of walk-outs. How to test: 8-week pilot in 5 clinics logging declined estimates and Plan B outcomes. Need ≥25% acceptance to justify the pricing.
4. Assumption: PIMS integration is achievable without a blocking partner agreement. How to test: Build against the ezyVet public API in 2 weeks; separately, get a straight answer from Covetrus on AVImark third-party access terms.
5. Assumption: Liability framing holds — that professional-liability carriers and state boards view this as acceptable decision support. How to test: One consultation with a veterinary practice attorney and an AVMA PLIT conversation, before writing marketing copy.

Risk flags

1. Regulatory / liability risk: Clinical decision support in a licensed profession. If a deferred diagnostic is later implicated in a bad outcome, the product’s role will be scrutinized. Mitigated by human-in-the-loop approval and thorough record documentation — but this must be designed in from day one, not retrofitted.
2. Incumbent absorption: Scribenote (a16z, 6,000+ vets) or VetRec (YC) already sit in the exam room capturing the conversation. Adding “generate an alternative plan” is a natural adjacent feature for them and they have the distribution. The window is real but not indefinite.
3. Professional resistance to tiering: Today’s Veterinary Business has published directly against good/better/best presentation. If the profession reads this as “discount software” or as pressure to practice below standard of care, adoption stalls regardless of the economics. The clinical-sequencing framing is not marketing polish — it’s survival.
4. Platform dependency: Value is much higher with PIMS integration, and the major PIMS vendors (Covetrus, IDEXX) are large companies with their own software ambitions who control that access.
5. Measurement difficulty: Proving incremental revenue requires a counterfactual — you can’t easily show what the client would have done without Plan B. Churn risk at renewal if the owner can’t see the ROI in their own numbers.

14. Structured verdict

Score:                  76/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       Technical founder paired with a practicing DVM co-founder or
                        heavily-committed clinical advisor. Comfort with slow, trust-driven
                        professional sales is mandatory.
Time to revenue:        8–12 weeks (paste-in v1, pre-sold via declined-estimate audits)
Capital to launch:      $25–35K
Top 3 assumptions to validate first:
  1. Vets will present an AI-drafted alternative plan to a client — 15 DVM prototype
     sessions on real declined estimates, need ≥60% "would use as-is or lightly edited"
  2. Plan B converts declines — 8-week, 5-clinic pilot, need ≥25% acceptance rate
  3. Owners pay for recovered revenue — free declined-estimate audit on 25 clinics,
     need ≥6 converting to paid pilot
Kill criteria:
  - Abandon if <25% of declined estimates convert on a Plan B across a 5-clinic,
    8-week pilot — the core economic premise fails and no pricing fixes it
  - Abandon if <40% of DVMs in prototype testing would present the output to a client
    without heavy rewriting — the comfort barrier is not actually addressable by software
  - Abandon if a funded scribe incumbent (Scribenote, VetRec, HappyDoc) ships an
    equivalent alternative-plan generator before your v1 reaches 25 paying practices
  - Abandon if veterinary practice counsel or PLIT flags unacceptable liability exposure
    that human-in-the-loop approval does not resolve

15. Next step — 1-week validation sprint

• Day 1–2: Build the paste-in prototype only — no integration, no auth, no billing. Paste an estimate plus signalment and presenting complaint, get a clinically-sequenced Plan B with per-item deferral rationale and citations. Assemble 12 realistic declined-estimate scenarios with a DVM advisor.
• Day 3–4: Recruit 15 practicing DVMs through veterinary Facebook groups, VIN, and early-career networks like MentorVet. Offer $75 for a 30-minute session. Show each 4 scenarios. Ask the single decisive question: “Would you hand this to a client right now?” Force a three-way answer — as-is / with edits / no. Capture verbatim objections; the objections matter more than the score.
• Day 5: Cold-email 40 independent practice owners from state licensing directories offering the free declined-estimate audit. Measure reply rate and how many will actually export the data.

Falsifiable decision rule: Proceed only if ≥9 of 15 DVMs say they’d present the output as-is or with light edits, and ≥6 of 40 practice owners agree to the audit. Below either threshold, the binding constraint is trust rather than tooling — and that is not something a solo builder fixes with more software.