RxIntact — case-intake validator for dental labs

1. One-liner

Reads a scrawled paper or photo dental Rx, catches missing shade/tooth/margin fields, and clears the case before it ships.

2. Trend signal — why now?

Three things converged in the last 12 months.

First, the pain is measured, not vibes. Industry write-ups and an NADL-cited figure put 34% of paper prescriptions arriving at the lab with at least one critical field missing or illegible, and a British Dental Journal audit found two-thirds of lab prescriptions non-compliant with basic completeness guidelines. A lab technician who has to chase those gaps loses 45–150 minutes a day to phone tag — the trade press frames it as “losing a full-time employee’s production to phone tag” over a year.

Second, the labor context makes that waste unbearable. There are over 6,000 unfilled dental-technician positions in the US and fewer than 15 accredited training programs left. Skilled tech-hours are the scarcest thing a lab has, and today they’re being spent reading someone else’s handwriting.

Third, the capability to fix it only recently got good and cheap. Vision LLMs in 2025–26 read messy handwritten forms, mixed print, and phone photos well enough to extract tooth numbers (Universal/Palmer/FDI), shade codes (VITA, 3D-Master), material, margin design, and pontic type — and, crucially, to flag what’s missing or ambiguous rather than silently guess.

Provenance:

• Signal 1 (Demand): “34% of paper prescriptions arrive at the lab with at least one critical field missing or illegible”; technician loses 45–150 min/day chasing gaps — https://trazalab.com/blog/dental-lab-clinic-communication-guide.html — 2026-07-16
• Signal 2 (Economic): US dental-lab market $7.6B, ~4,375 lab businesses, 60M restorations/yr, 4% remake rate at ~$420/remake; 6,000+ unfilled technician roles — https://www.ibisworld.com/united-states/industry/dental-laboratories/4087/ , https://www.speareducation.com/resources/spear-digest/the-cost-of-laboratory-remakes/ — 2026-07-16
• Signal 3 (Feasibility): Incumbent lab software (LabStar/Evident/3Shape LMS) manages cases after entry and targets 15+ tech labs at $99–250/mo; none auto-reads/validates the inbound paper or photo Rx — https://trazalab.com/best-dental-lab-software-small-labs.html — 2026-07-16 Category: Workflow automation

3. The opportunity

Every case a lab produces starts with a prescription. For the long tail of dental offices — and there are tens of thousands still not on a digital scanner — that Rx is a paper slip faxed, scanned, or photographed on a phone. The lab’s front desk keys it into their lab management system by hand. When a field is blank, illegible, or contradictory (shade says “A2” but photo shows a bleached prep), the case either stops for a callback or, worse, proceeds on a guess and comes back as a remake.

The incumbents — LabStar, Evident, 3Shape LMS — are case-management systems. They’re the system of record after a clean case exists. They assume someone already typed in correct data. They do nothing about the garbage-in problem at the door, and they’re priced and shaped for mid-to-large labs (15+ techs), leaving the thousands of 2–10 person labs underserved on exactly this.

The wedge: a focused tool that sits at intake, ingests the paper/photo/PDF Rx, extracts and validates every required field against the tooth number, and won’t let a case start until it’s complete — or auto-drafts the exact clarification message to send the office. AI is load-bearing: without vision-model extraction and field-validation, this is just another form.

4. Target market

• Primary customer: Owner or lab manager of an independent US dental laboratory, 2–15 employees, doing crown-and-bridge / removable / implant work, receiving a meaningful share of Rx as paper, fax, or phone photo. The person who signs off on remakes and hears the technicians complain.
• Why they buy (their words): “Missing shades. Unclear product selections. No margin details.” “Cases get paused mid-production because something important was left off the Rx form.” “The technician calls the office, gets put on hold, reaches front-desk staff who doesn’t know the clinical answer, leaves a message, waits for a callback that may come tomorrow.” Every one of those is a stalled case, an idle skilled tech, and a remake risk.
• Rough TAM reasoning: ~4,375 lab businesses (IBISWorld) up to ~27,000 counting every registered lab; the sweet spot of small independents is easily 8,000–12,000 labs. At $150–300/mo that’s a $15–40M niche — too small for the incumbents to chase hard, right-sized for a bootstrapper.
• Why now for them: The technician shortage means an hour of tech time reading handwriting is an hour of restorations not made. Labs are actively hunting for anything that reclaims skilled-labor hours in 2026.

5. Product sketch (MVP)

• Drop-in intake inbox: forward a fax/email, upload a scan, or snap a phone photo of the paper Rx.
• AI extracts structured fields: patient/case ID, tooth number(s), restoration type, material, shade, margin design, pontic, due date, special instructions.
• Completeness + sanity validation: flags missing required fields, illegible fields, and internal contradictions (shade vs. attached photo, tooth number vs. restoration type).
• One-click “clarification draft”: auto-writes the exact message to the dental office listing only what’s missing — no more “call the office and hope.”
• Clean, confirmed case export to whatever the lab already uses (LabStar/Evident/CSV/print) — we are the front door, not a rip-and-replace.
• Per-office Rx profile: learns each referring dentist’s handwriting quirks and abbreviations over time.
• Dashboard: cases held for clarification, average time-to-clear, and a running count of “remakes likely prevented.”

6. AI angle — what’s load-bearing

Remove the AI and there is no product. The entire value is (a) reading arbitrary handwritten/photographed Rx forms — every dentist’s slip and handwriting differs — into structured fields, and (b) reasoning about completeness and contradiction against dental domain rules (a #8 can’t be a “pontic,” an implant case needs an abutment spec, a shade must be a valid VITA/3D-Master code). That’s vision + domain-constrained validation, not OCR-plus-a-form. A dumb template would just capture the same blanks the paper had.

7. Localization angle (if any)

N/A — this is a US-first play. The wedge is US shade systems, tooth-numbering conventions (Universal), and the specific incumbent gap in the US small-lab market. The same problem exists in EU/UK (FDI numbering) and India, and those are clean expansion markets, but forcing a localization angle at launch would dilute focus. Start US, follow the paper.

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

• Pricing: $199/mo flat per lab for up to 200 Rx/mo, $349/mo for higher volume. Priced against a single prevented remake ($420) and against reclaiming ~1 FTE-equivalent of chased-info time — trivially ROI-positive.
• ACV: ~$2,900/yr blended.
• Rough math to $1M ARR: ~345 labs × $2,900 = $1.0M. That’s ~3–4% of the addressable small-lab pool.
• Rough math to $5M ARR: ~1,150 labs (~10–14% of the pool) plus a usage-based tier for high-volume labs and a per-seat add-on for multi-location groups. Requires expanding beyond intake into light case-tracking so it’s stickier than a single workflow.
• Expansion path: start at intake validation → add the clarification-messaging loop (two-way with offices) → add a lightweight portal dentists use to submit clean digital Rx (flips the model from “we clean your mess” to “we prevent it”), which grows ACV and defensibility.

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

• Dental Lab Network + NADL channels: the forum (dentallabnetwork.com) and the National Association of Dental Laboratories are where lab owners argue about software daily. Post a teardown of the “34% incomplete Rx” data with a before/after of a real slip → structured case. Offer 20 labs a free 30-day “we’ll process your worst 50 Rx” pilot.
• Cold outreach to the directory: scrape the NADL / state dental-lab association member lists (thousands of named labs with contact info). Send a 60-second Loom: photograph a deliberately-messy Rx, run it through the tool live, show the missing-field flags. Expect a strong reply rate from a group whose #1 complaint this directly hits.
• Trade shows & regional meetings: LMT LAB DAY (Chicago) and regional lab guild meetings are dense with exactly these buyers. A live “snap-your-Rx” demo booth converts because they feel the pain weekly.
• Referral loop via prevented remakes: the dashboard’s “remakes likely prevented” number is a shareable proof artifact — labs brag about it to peers.

10. Build complexity — justification

Low. Off-the-shelf vision LLM for extraction, a domain rule-set for validation (tooth-number/shade/material logic — buildable with a lab-tech advisor, no custom model), standard web app, an email/fax ingest, and CSV/print export. No hardware, no ML training, no regulatory approval (this is a workflow tool, not a medical device — it doesn’t diagnose or design). A solo technical founder with a part-time dental-lab domain advisor ships a credible v1 in 8–12 weeks. Deepest work is the domain rule-set and per-office handwriting adaptation, both incremental.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	Workflow/data-entry tool; not a medical device, doesn’t design restorations. Standard PHI handling (BAA, encryption) required.
Ethical — no harm / dark patterns	✅	Reduces remakes and errors; strictly beneficial. Human confirms every case before it proceeds.
Market exists (evidence above)	✅	34% incomplete Rx, $7.6B market, quantified time loss, active software spend.
1–5 person team can build this	✅	Solo + advisor, 8–12 weeks to v1.
Launchable with <$50K / ₹40L	✅	Off-the-shelf APIs; main cost is inference + a domain advisor.

All five pass.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	16/20	Hair-on-fire for the tech, but the owner tolerates it today; 45–150 min/day/tech and $420 remakes make it a real, regular bleed.
Demand evidence	15	12/15	Hard numbers (34% incomplete, ⅔ non-compliant BDJ audit, quantified time loss) + funded incumbents nearby. Missing: verbatim “I’d pay for this” from owners.
Build feasibility	15	13/15	Off-the-shelf vision + a rule-set; the domain logic is the only real work.
Distribution clarity	15	11/15	Named channels (NADL, DentalLabNetwork, LMT LAB DAY) and a scrapeable directory, but conversion on a sleepy, traditional buyer is unproven.
Revenue mechanics	15	11/15	ROI is obvious vs. one remake; risk is a small pool and whether flat $199 sticks vs. usage.
Time to first revenue	10	8/10	Pilot-to-paid in weeks; buyers feel the pain immediately.
Defensibility	10	5/10	Execution + accumulating per-office Rx profiles and the office-side portal, but the extraction itself is copyable.
Total	100	76/100

13. Qualitative modifiers

Founder-fit tags

technical-heavy · domain-expertise-required — needs a builder comfortable with vision-model pipelines and a dental-lab advisor to nail the validation rules and buyer language.

Key assumptions to validate (3–5)

1. Assumption: A meaningful share of small labs’ inbound Rx is still paper/photo (not already digital). How to test: Ask 25 lab owners what % of their Rx arrive as paper/fax/photo; need median ≥30%.
2. Assumption: Owners will pay ~$199/mo to reclaim tech-hours and cut remakes. How to test: Pre-sell 10 pilots at $199/mo with a 30-day out; ≥5 convert to paid.
3. Assumption: Vision extraction hits usable accuracy on real-world messy slips. How to test: Run 200 real anonymized Rx; measure field-level extraction + missing-field-flag precision; need ≥90% on flagging.
4. Assumption: Labs will let us sit at intake without ripping out LabStar/Evident. How to test: Confirm clean export/handoff satisfies 8/10 pilot labs.

Risk flags

1. Market timing / substrate erosion: The whole wedge is paper Rx. As digital scanners spread, the paper pool shrinks — but slowly, and the office-side portal is the hedge that turns this from a decliner into a preventer.
2. Small TAM: 8–12K labs caps this well under $10M ARR without adjacent expansion (EU/UK, dentist-side portal, other lab types). Fine for a bootstrapper, not a VC play.
3. PHI / compliance overhead: Handling patient-linked prescriptions means BAAs and proper data handling from day one — table stakes, but real work.
4. Incumbent yawn-then-copy: If a lab owner asks LabStar/3Shape for this, they could bolt on an intake reader. Head start + office-side portal + per-office learning is the only moat.

14. Structured verdict

Score:                  76/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       Technical founder (vision-LLM comfortable) + part-time dental-lab-tech advisor
Time to revenue:        6–10 weeks (pilot-to-paid)
Capital to launch:      $8–15K ($ for inference, BAA/legal, advisor)
Top 3 assumptions to validate first:
  1. ≥30% of small-lab inbound Rx is still paper/photo — survey 25 owners
  2. ≥5/10 pilots convert to paid at $199/mo — pre-sell pilots
  3. ≥90% precision flagging missing fields on 200 real messy slips — accuracy test
Kill criteria:
  - Abandon if <30% of surveyed labs' Rx volume is non-digital (substrate too thin)
  - Abandon if missing-field flag precision stays <85% after tuning on 200 real slips
  - Abandon if <3/10 pilots convert to paid after a 30-day free run

15. Next step — 1-week validation sprint

• Day 1–2: Pull the NADL / state-association member directory. Cold-call/DM 25 small-lab owners: ask (a) what % of Rx is paper/photo, (b) minutes/day lost to chasing missing fields, (c) would they pay $199/mo to fix it.
• Day 3–4: Collect 200 real (anonymized) messy Rx from 3 friendly pilot labs. Run them through an off-the-shelf vision model with a first-pass validation prompt. Measure field extraction + missing-field-flag precision.
• Day 5: Decide. Go if ≥30% median paper share AND ≥90% flag precision AND ≥5 owners verbally commit to a paid pilot. Otherwise, no-go — the substrate or the accuracy isn’t there.

Falsifiable outcome: a precision number on 200 real slips and a count of verbal paid-pilot commits. No vibes.