DenialWorth

1. One-liner

Ranks a dental practice’s denied claims by what they’re actually worth chasing, so the winnable ones stop being written off.

2. Trend signal — why now?

One statistic carries this entire idea: 67% of denied dental claims are never resubmitted. Not lost on appeal — never attempted. Between 15% and 20% of dental claims are denied on first submission, and denials cost practices somewhere in the range of $50,000 to $120,000 annually, or 5–10% of revenue.

Read that carefully, because it tells you the problem is not what everyone thinks it is. This is not money that got fought over and lost. This is money that was earned, delivered, adjudicated, and then quietly abandoned because nobody in the front office had time to decide whether it was worth twenty minutes.

The second signal explains why now. A survey of 160+ dental billing professionals found 78% report increased claim denials and payer scrutiny over the past 12 months — and critically, the survey attributes the primary cause to “shifting insurance policy interpretations around medical necessity and frequency limitations, not billing errors.” That distinction is the whole product. The existing tool category — claim scrubbers, coding validators, eligibility checkers — is built to catch your mistakes. But the denials growing fastest aren’t your mistakes. They’re a payer reinterpreting a frequency limitation this quarter in a way they didn’t last quarter. A scrubber cannot see that coming, and it cannot tell you whether to fight it.

Third: the labor to do this by hand is evaporating. 71% of billing professionals cite insurance verification as their top daily challenge, roughly 90% of dental practices are still struggling to hire, and labor shortages have driven an estimated 11% reduction in practice capacity. The same survey names the core condition the “Efficiency Paradox” — “strong performance sustained by unsustainable manual effort.” When the front office is underwater, denial triage is the first thing dropped, because it’s the only task with no patient standing in front of you.

And the clock is real. Delta Dental appeals must be filed within 180 days of the denial notice; Cigna also allows 180 days. Abandoned denials don’t stay winnable — they expire silently.

Provenance:

• Signal 1 (demand): 67% of denied dental claims never resubmitted; denials cost practices $50K–$120K/yr and 5–10% of annual revenue; 15–20% first-pass denial rate — https://hellopearl.com/blog/13-tips-to-reduce-dental-insurance-denials-improve-collections-pearl-ai — 2026-07-20
• Signal 2 (feasibility): Open Dental exposes a REST API where “developers don’t need approval to read data” with “no enrollment fee”; Dentrix Ascend exposes REST endpoints including insurance claims — a per-claim triage read is buildable without a partnership gate — https://mentera.ai/blog/dental-pms-comparison-ai-layer-integration — 2026-07-20
• Signal 3 (economic): 2026 survey of 160+ dental billing professionals — 78% report rising denials driven by shifting policy interpretation “not billing errors”; 71% cite insurance verification as top daily challenge; 58% have adopted or plan to adopt AI/automation in 2026 — https://www.zentist.io/blog/blog-2026-dental-rcm-trends — 2026-07-20 Category: Workflow automation

3. The opportunity

The dental RCM market sells two things, and neither one solves this.

Option A: claim scrubbers and eligibility tools. These run before submission and catch formatting and coding defects. Useful. But 78% of the denial growth is policy reinterpretation, which by definition passes every scrubber — the claim was clean, the payer just decided differently. Post-denial, the scrubber has nothing to say.

Option B: outsourced billing companies, charging 4–10% of collections. That pricing is the tell. A practice collecting $900K pays $36K–$90K a year, levied on all collections — including the 80–85% of claims that sailed through untouched. You’re paying a percentage of money you would have received anyway to get help with the sliver that went wrong. For a 1–4 chair practice, that math is why so many never outsource and just absorb the write-offs instead.

The gap between them is the actual job: a denied claim landed, and someone has to decide in the next thirty seconds whether it’s worth twenty minutes. Today that decision is made by an overloaded front-office coordinator using gut feel, and 67% of the time the answer is “no” by default — including on claims that would have been paid on a one-page appeal.

That’s a ranking problem, not a writing problem. And it’s why I’m not building another appeal-letter generator. Drafting the letter is the easy, commoditized half. Knowing which fifteen of this month’s sixty denials deserve a letter — based on this payer’s actual behavior on this denial code at this practice — is the half nobody sells to a solo practice.

The wedge: make the write-off decision explicit and evidence-backed instead of implicit and exhausted. Then draft only the ones that clear the bar.

4. Target market

• Primary customer: Owner-dentist or office manager at an independent 1–4 chair general practice in the US, doing $600K–$1.5M in annual collections, with 1–2 front-office staff and no dedicated billing specialist. Not DSOs — they have RCM teams and Zentist-tier tooling already.
• Why they buy: They know they’re leaving money on the table and can’t quantify it. The pain is specific and recurring: a denial EOB arrives, the coordinator glances at it, doesn’t recognize the denial reason, has three patients checking in, and files it in the “deal with it later” pile that never gets dealt with. At 15–20% denial rates, a practice submitting 200 claims/month is generating 30–40 denials a month — and abandoning roughly two-thirds of them.
• Rough TAM reasoning: ~180,000+ dentists in active US practice, the large majority still in independent or small-group settings despite DSO consolidation. Even restricting to solo/small general practices with meaningful insurance participation, the reachable base is comfortably in the tens of thousands. I need 400 of them.
• Why now for them: Denials are up (78% report it), staff to handle them is unavailable (90% struggling to hire, 11% capacity reduction), and patient out-of-pocket costs are rising — 31% named that the trend most likely to impact their 2026 business — which makes insurance dollars that should have been collected more painful to lose.

5. Product sketch (MVP)

• Denial inbox with a worth-chasing score. Every denied claim from the practice management system, ranked 0–100 on expected recoverable value: dollar amount × estimated win probability for that payer + denial code combination, with days-to-deadline as a decay factor.
• A plain-English reason for each score. “Delta Dental, D4341, frequency limitation — you’ve won 4 of 5 of these with perio charting attached. $340. 162 days left.” Never a bare number; the coordinator has to trust it in five seconds.
• The write-off list, stated out loud. The bottom of the ranking is a defensible “don’t chase these” list with reasoning. Giving someone permission to stop is as valuable as telling them to start, and it’s what makes the top of the list credible.
• One-click appeal packet for anything above the bar — appeal letter, the payer’s own policy language on that denial reason, and a checklist of the attachments that specific payer wants (narrative, radiographs, perio chart).
• Deadline countdown per claim, against the 180-day appeal windows, with escalating alerts as claims approach expiry.
• Monthly recovered-revenue report — dollars recovered, dollars still winnable, dollars expired unclaimed. This is the retention mechanic and the renewal argument in one artifact.
• Payer pattern digest. “Cigna started denying D0180 at a higher rate 6 weeks ago” — the shifting-interpretation signal that no scrubber catches.

6. AI angle — what’s load-bearing

Two jobs, and the product collapses without either.

Reading unstructured denial reasoning. Denial reasons arrive as terse, payer-specific, non-standardized text and remark codes across ERAs and paper EOBs. Mapping “processed per contractual frequency provisions” to this is a frequency-limitation denial on a code where this payer has historically paid on appeal with charting attached is exactly the semantic normalization LLMs are good at and rules engines have failed at for twenty years.

Estimating win probability with almost no data at first. Cold-start is the hard part and I’d rather name it than hide it. Practice one has no appeal history, so v1 leans on the payer’s own published policy language plus denial-reason semantics to produce a calibrated prior. As claims resolve, outcomes feed back — per payer, per code, per practice. By month twelve, the estimate is grounded in the operator’s own realized outcomes across thousands of adjudicated claims.

Remove the AI and you have a spreadsheet of denials sorted by dollar amount — which is roughly what practices already fail to use. The ranking is the product.

One deliberate constraint: the product never auto-submits. It ranks and drafts; a human sends. This keeps it clear of practicing billing on the customer’s behalf and keeps the liability posture clean.

7. Localization angle (if any)

N/A — this is a US-only play. The entire value sits in US dental payer behavior: CDT codes, ERA/EOB formats, payer-specific appeal windows, and the specific structure of US dental insurance with its annual maximums and frequency limitations. None of that transfers. A UK or India version would be a different product against a different payer structure, not a translation.

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

• Pricing: $299/month per practice location, flat. Deliberately not a percentage of collections — that’s the incumbent model and its weakness. A flat fee means the practice keeps 100% of what it recovers, which makes the ROI argument trivial: recover one $340 claim a month and it’s paid for.
• ACV: ~$3,600.
• Rough math to $1M ARR: 280 practices × $299 × 12 ≈ $1.0M.
• Rough math to $5M ARR: ~1,400 practices, or a blend — roughly 900 single locations plus 80 small groups on a multi-location tier at $1,200–2,500/mo. Small groups (3–10 locations) are the natural expansion; they have the same problem with more zeros and still no RCM department.
• Expansion path: Per-location pricing scales with group growth. Natural upsells: pre-submission risk flagging for the same payer patterns (using the accumulated outcome data in the other direction), and a fee-schedule/underpayment check — the same ERA data supports detecting claims paid below contracted rate, which is an adjacent leak.

Gross margin is high — the per-claim inference cost is cents against a $299 seat.

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

• Open Dental first, deliberately. It’s the one major PMS with an open REST API, no approval gate, and no enrollment fee. That means I can build and demo without a partnership negotiation, and Open Dental practices skew independent and price-conscious — exactly my buyer. Open Dental’s user community and its annual meeting are a concentrated, reachable population.
• The free denial audit as the entire sales motion. Practice exports 12 months of denied claims (a standard report in every PMS); I return a one-page number: “$47,300 in denials, $19,800 of it still inside the appeal window, here are the 23 claims I’d chase first.” That is a specific dollar figure about their own money, produced before they pay anything. Target 200 audits to convert 100 customers — and the audit doubles as the training data that sharpens the model.
• Dental billing consultants and fractional office managers as channel. These people serve 5–20 practices each and are drowning; a tool that tells them where to spend their hours makes them more profitable. Revenue-share or straight referral fee. Twenty consultants is a few hundred practices of reach.
• Facebook groups and the Dentaltown forum. Independent dentists and office managers congregate in large, active private groups where practice-management complaints are the daily content. Post the aggregate finding — “here’s the denial-abandonment rate across 40 practices we audited” — not a pitch. Original data earns the audience; the audit offer converts it.
• Study clubs and state dental association meetings. Local, cheap, high-trust, and the audit demo lands in person in ten minutes.

10. Build complexity — justification

Low. Off-the-shelf on every axis: read claims and denial data via the Open Dental REST API (and ERA 835 files, a stable published format), LLM for denial-reason normalization, a scoring model that starts as a transparent heuristic and only becomes learned once real outcome data exists. No custom ML infrastructure on day one, no hardware, no novel research.

The genuine work is unglamorous: denial-code and payer-behavior mapping, and handling the variability of real ERA files. Call it 8–10 weeks to a demoable v1 for a technical founder, with the first version’s “model” being a defensible heuristic — which is honest, because with zero customers there is no outcome data to learn from anyway.

Second PMS integration (Dentrix Ascend, then Eaglesoft via Patterson Innovation Connection) is the main scaling cost and can wait until Open Dental traction is proven.

11. Gating checklist

Gate	Pass?	Note
Legal in target market	✅	Practice’s own claims data, accessed with their authorization via a sanctioned API. Requires a BAA and HIPAA-compliant handling — standard, well-trodden, not a licensing barrier. Product never auto-submits to payers.
Ethical — no harm / dark patterns	✅	Helps practices collect money they legitimately earned for care already delivered. Explicitly recommends abandoning weak claims rather than spamming payers with appeals. No patient is billed more.
Market exists (evidence above)	✅	15–20% denial rate, 67% never resubmitted, $50K–$120K annual exposure, existing outsourcers charging 4–10% of collections.
1–5 person team can build this	✅	Low complexity, one open API to start, 8–10 weeks to v1.
Launchable with <$50K / ₹40L	✅	Solo founder, inference costs, HIPAA-compliant hosting, a BAA and a lawyer review. Well under.

All five pass.

12. Feasibility score

Axis	Weight	Score	Notes
Problem intensity	20	17/20	Quantified, recurring, monetary — $50K–$120K/yr abandoned. Felt monthly (30–40 denials). Not quite hair-on-fire only because it’s a silent loss: practices tolerate it precisely because it never shows up as a bill. That invisibility is also the sales obstacle.
Demand evidence	15	13/15	Multiple independent sourced signals: 67% non-resubmission, 78% reporting rising denials, 71% naming verification the top challenge, and a live priced market (outsourcers at 4–10% of collections). A skeptic nods.
Build feasibility	15	12/15	Open API, published ERA format, off-the-shelf models, 8–10 weeks. Docked for real-world ERA messiness and the fact that PMS #2 and #3 are meaningful additional work.
Distribution clarity	15	12/15	The free denial audit is a strong, concrete, self-qualifying wedge that produces a dollar figure about the prospect’s own money. Docked because dentists are a famously noisy, heavily-marketed-to audience and Open Dental’s share limits the beachhead.
Revenue mechanics	15	12/15	$299/mo flat against a $50K–$120K problem is an easy ROI story, and 280 practices to $1M is reachable. Docked because SMB dental churn is real and $299 is a considered purchase for a 1-chair practice.
Time to first revenue	10	8/10	Audit-to-paid can close in weeks; the audit itself is deliverable before the product is finished. Not a 9–10 because HIPAA/BAA paperwork adds friction to the first close.
Defensibility	10	4/10	Honest score. The wedge is copyable — the moat is the accumulating payer-outcome dataset, which is worth little at month 3 and meaningfully compounding by month 12. Execution-and-data moat, nothing more.
Total	100	78/100

13. Qualitative modifiers

Founder-fit tags

technical-heavy · domain-expertise-required

You need someone who can build the integration and someone who genuinely understands dental billing. If the founder isn’t the latter, a dental billing consultant as advisor or co-founder is non-optional — the win-probability priors and denial-code mapping are where domain knowledge decides whether the ranking is credible or garbage.

Key assumptions to validate (3–5)

1. Assumption: The 67% non-resubmission rate holds at 1–4 chair independent practices specifically (the aggregate may be skewed by large groups). How to test: Run 20 free denial audits and measure actual resubmission rates in the raw exported data. This is directly observable, not a survey question.
2. Assumption: A ranked list changes behavior — coordinators actually chase the top-ranked claims rather than ignoring a new dashboard like they ignored the old pile. How to test: In a 10-practice pilot, measure resubmission rate before vs. after over 60 days. If it doesn’t move, the product is a report, not a tool.
3. Assumption: Win probability is predictable enough from payer + denial code + attachments to beat coordinator gut feel. How to test: Backtest against 12 months of historical denials with known outcomes from 10 practices. If the ranking can’t beat “sort by dollar amount,” there’s no product.
4. Assumption: Owner-dentists will pay $299/mo flat rather than demand contingency pricing. How to test: Present both to 30 practices in the audit follow-up; measure preference and close rate.

Risk flags

1. Platform dependency: Open Dental’s API is the beachhead. If access terms change, or if Dentrix/Eaglesoft integrations prove gated behind expensive partner programs, the addressable market shrinks hard. Mitigate by supporting ERA 835 file upload as a PMS-agnostic fallback path from day one.
2. Incumbent absorption: This is a plausible feature for Dentrix, Zentist, or a large billing company to ship. The 4/10 defensibility score reflects that. Speed and the outcome dataset are the only defenses; a 12-month head start on payer-outcome data is the realistic goal.
3. Compliance overhead: PHI means HIPAA, BAAs, and a security posture that a consumer-SaaS founder may underestimate. Not a blocker — it’s a well-trodden path — but it’s real cost and it slows the first close.
4. The invisibility problem: Practices have tolerated this loss for years precisely because it never appears on a statement. Selling against an unfelt loss is harder than selling against a bill. The free audit exists specifically to convert the invisible into a number, and if the audit doesn’t shock people, the whole GTM thesis weakens.

14. Structured verdict

Score:                  78/100
Verdict:                GO
Confidence:             Medium
Best-fit builder:       Technical founder paired with a dental billing consultant as
                        co-founder or advisor. Integration work is routine; the
                        domain judgment is what makes the ranking trustworthy.
Time to revenue:        8–12 weeks (audits can be sold before v1 ships)
Capital to launch:      $15–25K (HIPAA-compliant hosting, inference, legal/BAA review)
Top 3 assumptions to validate first:
  1. 67% non-resubmission holds at 1-4 chair practices — measure directly in 20 free audits
  2. Ranking beats sort-by-dollar-amount — backtest on 12 months of resolved denials
  3. Ranked list actually changes resubmission behavior — 10-practice, 60-day pilot
Kill criteria:
  - Abandon if fewer than 15 of 20 audited practices show >40% of denials abandoned
    (the core premise is then wrong)
  - Abandon if the ranking cannot beat naive dollar-sort on backtested win rate
  - Abandon if pilot practices' resubmission rate doesn't rise materially in 60 days
  - Abandon if Dentrix and Eaglesoft integration paths both prove closed AND the
    ERA-upload fallback shows unacceptable data quality

15. Next step — 1-week validation sprint

• Day 1–2: Recruit 20 independent practices (Open Dental Facebook groups, Dentaltown, two local study clubs) for a free denial audit. The pitch is one sentence: “Send me your last 12 months of denied claims and I’ll tell you how much is still collectible.” No product required — this is done by hand.
• Day 3–4: Manually analyze the exports. Measure the two numbers that matter: what percentage of denials were never resubmitted, and what dollar value sits inside a still-open 180-day appeal window. Hand-rank the top 20 claims per practice, with an experienced dental biller reviewing the ranking.
• Day 5: Deliver each audit and immediately ask for $299/mo, framed against the specific figure. Decide on falsifiable outcomes.

Go / no-go: Go if (a) median abandonment rate across the 20 practices exceeds 40%, (b) median still-collectible-in-window value exceeds $8,000, and (c) at least 5 of 20 practices commit to paying $299/mo on the spot. Fewer than 3 commitments and the loss is real but not felt keenly enough to sell against — which is the specific way this idea dies.