Underpay Scout
Upload EOB PDFs and your contract fee schedule; Underpay Scout finds line-item underpayments and generates ready-to-send dispute packets in minutes.
One product, one founder, thirty days. Everything below is built to be executed inside the contest window — the official rules still decide what counts.
Decide in 20 seconds
The facts that decide whether this is worth thirty days of yours.
Billing manager or practice administrator at a 1-10 provider US specialty practice (dermatology, orthopedics, gastroenterology) who posts EOBs and manages payer contracts.
Operational simplicity 1.9/10
Remove regulated data, human review steps and custom infrastructure; keep the product self-serve and stateless on managed services.
Why this could work in 30 days: one founder can ship the MVP in 4 days; the plan targets a first sale on day 12; 3 sales at $499 clears $1,000 of collected revenue; strongest signal is speed to first revenue at 7.3/10.
Score and projections come from the generated report and are not guarantees. The $100K Hackathon is run by Ship or Die — read the official rules before you build.
Your first 24 hours
11 concrete actions. Tick them off as you go — nothing is saved or sent.
Product opportunity and positioning
Target customer
Billing manager or practice administrator at a 1-10 provider US specialty practice (dermatology, orthopedics, gastroenterology) who posts EOBs and manages payer contracts.
Why now
The buyer already owns every input, so the first dollar of value appears on the first upload with no EHR integration, no clearinghouse migration and no service takeover.
Problem
EOBs get posted line by line but never audited against contracted rates, so sub-$100 underpayments are silently absorbed and cumulative leakage is invisible.
Solution
Buyer uploads remittance PDFs plus a contract fee schedule PDF/CSV; the product extracts claim lines, matches CPT/modifier to contracted rate, computes expected-vs-paid variance, and outputs a per-payer dispute packet (PDF letter, line-item CSV, evidence appendix).
1. Product opportunity and positioning
Underpay Scout audits claims that were already paid. The buyer posts an EOB, accepts the allowed amount as truth, and moves on; nobody re-reads the contract fee schedule for line 14 of a 60-line remittance. That is where sub-$100 underpayments hide, and because each one is small, no one ever builds a case around them. Underpay Scout builds the case automatically: remittance lines in, contract variance out, plus a packet the practice can actually send to the payer.
The buyer is one person: the billing manager or practice administrator at a 1-10 provider US specialty practice who posts EOBs and holds the payer contracts. The trigger is concrete: month-end close, a contract renegotiation, or a new manager told to find revenue without new headcount. Those buyers answer LinkedIn messages. They do not run procurement.
JEV scored this concept 53.3/100 (band C, confidence 0.62). The strongest criteria are speed to first revenue (7.3 normalized) and reachable customers (6.33); the weakest are operational simplicity (1.91) and market demand (4.18). That profile dictates the plan: sell fast with a credible free audit, and keep the regulated surface deliberately small. Every build decision below reduces PHI exposure and manual labor rather than adding capability.
Positioning against the three adjacent options is clean. A billing service takes a percentage of collections and wants the whole revenue cycle; Underpay Scout takes one flat fee and wants one upload. A clearinghouse submits claims and has no opinion about contracted rates. Denial-appeal tools chase claims that were rejected, not claims that were paid short. Underpay Scout sits in the gap between posting and writing off, and its output is a dispute packet, not an appeal letter.
The reason this is sellable in 30 days is the demo format: the founder runs 20 claims for free and returns a number. If that number is $0, the prospect loses ten minutes. If it is $700, the $499 packet sells itself. No integration, no pilot, no security review of a production system. The audit is the pitch.
2. Smallest sellable MVP
One upload flow and one output. The buyer drags remittance PDFs and a fee schedule into the browser, sees a redaction preview, and gets a variance table plus a downloadable packet. Anything that does not contribute to that sentence is out: no EHR connection, no clearinghouse, no automated payer portal submission, no denial appeals, no patient billing, no RCM dashboard, no 24/7 support promise. Manual override for the top CPT rates is in scope because fee schedules are messy and the audit must be defensible. Processing transparency (a visible audit log of what was parsed) is in scope because trust is the real blocker at first sale.
3. Technical implementation plan
Next.js on Vercel, Postgres for accounts and job state, S3 for short-lived document storage, pdf.js for client-side preview, GPT-4o for remittance and fee-schedule extraction plus letter drafting, Stripe for the single $499 checkout, Resend for packet delivery. Four build days: extraction spike, rate ingestion, variance engine, packet plus checkout. The known break point is remittance format variance; the mitigations are a small template library per payer, per-field confidence flags, and a manual correction cell in the variance table so a bad parse is fixed in seconds instead of failing the job.
4. Pricing and monetization
One product: Underpayment Audit Pack, $499 one-time, up to 1,000 claim lines, per-payer packet, variance report, one revision by email. The price is high enough to signal a real audit and low enough to clear without procurement, and one recovered line often exceeds it. Delivery cost per pack is a few dollars of extraction tokens and storage, so margin holds at every volume. Repeat revenue comes from additional packs at the same price for practices with more volume; no subscription is introduced during the contest.
5. First-customer acquisition
The channel is LinkedIn direct outreach to billing managers and practice administrators at 2-10 provider specialty practices, 40 messages a day, with a free 20-claim audit as the offer. Each audit returns a redacted sample packet and a found-money number. Supporting surfaces, not replacements: three medical billing Facebook groups and ten local practices within driving distance. No paid ads, no SEO, no partnership.
6. 30-day roadmap shape
Days 1-5 build and ship the MVP (payment rails connected on day 1, before any code). Days 6-7 harden accuracy and produce sales assets. Days 8-14 are list build, first 120 outreach sends, five free audits and the first paid close targeted for day 12. Days 15-21 add message variants, local calls, repeat-pack offers and referral asks. Days 22-28 automate packet assembly, publish a trust page, clear open quotes. Days 29-30 are a final outreach wave and collection, with every payment landing before December 16, 13:00 UTC.
7. Revenue and profit scenarios (projections only)
Three projections, none verified: 4 customers ($1,996 gross), 12 customers ($5,988 gross), 30 customers ($14,970 gross) at the $499 price. Processing at roughly 3.2% (Stripe 2.9% + $0.30). Base case carries one refund; upside carries two refunds, one chargeback and $300 of testing spend. Hackathon profit subtracts refunds, chargebacks, processing and ad spend only; business profit additionally subtracts $85-$320 of hosting, storage and extraction cost. These are hypotheses built on stated assumptions, not commitments or traction.
8. Risks and validation experiments
PHI handling is the top risk; it is tested in week one by confirming BAA-covered storage and model terms before taking a single real EOB. Extraction accuracy is tested by running ten public EOB formats and counting clean lines. Willingness to pay is tested by whether 40 messages produce at least two accepted free audits by day 10. Value size is tested by measuring found money per 1,000 lines across the first five audits; a median below $500 means the segment, not the price, is wrong.
9. First 24 hours
Connect Stripe and TrustMRR, scaffold and deploy the app shell, write the audit spec, then run a five-EOB extraction spike before sleeping. The day-one artifact is a live URL, a connected payment account, and claim lines in a database.
Smallest sellable MVP
A single upload flow: remittance PDFs + contract fee schedule in, a variance table and a downloadable per-payer dispute packet out, gated by one $499 Stripe checkout.
Must-have features
- Secure upload for EOB/remittance PDFs and fee schedule PDF/CSV
- Redaction preview shown before processing
- LLM extraction of claim lines: payer, claim ID, DOS, CPT, modifier, billed, allowed, paid, patient responsibility
- Contract rate table parsed from fee schedule with manual override for top CPTs
- Variance engine computing expected vs paid by claim line and payer
- Per-payer dispute packet: PDF letter, line-item CSV, evidence appendix with contract rate citations
- Dashboard with total underpayment found, packet download, audit history
- Stripe $499 checkout with token-based access
Explicitly out of scope
- EHR integration
- Clearinghouse or automated claim submission
- Denial appeals
- Patient billing and collections
- Full revenue cycle management
- Payer portal automation
- Subscription billing
- 24/7 support
Demo flow (what the buyer sees)
- Buyer opens link from outreach message
- Uploads two remittance PDFs and one fee schedule
- Sees redaction preview and confirms
- Job runs; variance table renders with flagged lines and confidence flags
- Buyer overrides two CPT rates in the table
- Clicks Generate Packet
- Downloads PDF letter, CSV and evidence appendix
- Pays $499 via Stripe for the full 1,000-line pack
Technical implementation plan
Stack
Architecture
Next.js front end handles upload, redaction preview and the variance table. An API route stores files in S3 under a per-job prefix, enqueues a processing job, and records status in Postgres. A worker calls GPT-4o in two passes: remittance extraction into a claim-line schema, and fee-schedule extraction into a CPT/modifier rate table. A deterministic variance engine matches lines to rates, computes expected minus paid, and flags anything above the threshold. The packet generator renders the letter and evidence appendix to PDF, writes the CSV, emails via Resend, and the job prefix is deleted on the retention schedule.
Build steps
- Day 1: Stripe and TrustMRR connected, Next.js shell deployed, audit spec written
- Day 2: claim-line schema, GPT-4o extraction on 5 sample EOBs, redaction preview
- Day 3: fee schedule parser plus manual override table keyed by payer and CPT
- Day 4: variance engine and dashboard table
- Day 5: packet generator, checkout, production deploy
- Day 6: confidence flags, manual correction cell, processing audit log
- Day 7: redacted sample packet, Loom walkthrough, offer page
- Day 8+: template library per payer as new formats appear in real audits
AI components
- GPT-4o vision/text extraction of remittance claim lines
- GPT-4o extraction of fee schedule tables into CPT rate records
- GPT-4o drafting of the dispute letter body from structured variance data
- No model output is used as the source of truth for a dollar figure; the variance engine computes it deterministically
Pricing and monetization
Underpayment Audit Pack
Entry offerFor Billing manager at a 1-10 provider specialty practice
- Up to 1,000 claim lines audited
- Per-payer dispute packet: PDF letter, line-item CSV, evidence appendix
- Downloadable variance report
- Email support for one revision
Model: Single one-time pack: $499 per audit of up to 1,000 claim lines, no subscription.
First revenue offer: Free audit of 20 claims from your own EOBs, returned as a redacted sample packet with a found-money number. If it finds more than $500, the full 1,000-line packet is $499 one-time.
Expected ACV: $499
Upsell path: Additional packs at the same $499 price for practices with more than 1,000 lines per month; a repeat-purchase path only, no subscription introduced during the 30 days.
First-customer acquisition
LinkedIn direct outreach to billing managers and practice administrators at 2-10 provider specialty practices, offering a free 20-claim underpayment audit.
Where to find them
- LinkedIn search: title Billing Manager or Practice Administrator, specialty dermatology, orthopedics, gastroenterology, 2-10 providers
- LinkedIn search: title Office Manager plus state, filtered to single-specialty practices
- Three medical billing and practice management Facebook groups
- Ten specialty practices within 30 minutes of the founder, contacted by phone or in person
- Public state specialty association member directories
Exact outreach motions
- Connection request plus one-line offer: free 20-claim underpayment audit, no upload of full records required
- Loom-first variant: 60-second video of a sample EOB becoming a variance table
- Follow-up 2 after four days with a redacted sample packet attached
- Facebook group post: offer the free audit, post the redacted packet as proof
- Local calls: ask for the billing lead by name and offer the same audit
- Referral ask to every buyer: one peer practice introduction for a free audit
Message template
Hi {first} - I built a tool that checks posted EOBs against your contracted rates and flags line-item underpayments, then writes the payer dispute packet. Want a free audit of 20 claims? You send two EOBs and your fee schedule, I send back a redacted packet and the dollar amount I found. No integration, no call required.30-day execution roadmap
Week 1 · Build & validate · 7 days
Payment path live before any customer exists
- Create Stripe account and connect it to TrustMRR
- Scaffold Next.js app on Vercel
- Write the 20-claim audit spec: inputs, outputs, acceptance
Deliverable: Live placeholder page and a connected payment account
Prove claim lines can be pulled from real EOBs
- Define the claim-line JSON schema
- Run GPT-4o extraction on 5 sample EOBs
- Build the PHI redaction preview screen
Deliverable: Claim lines extracted from 5 EOBs with CPT, allowed and paid captured
Turn fee schedules into a usable rate table
- Parse fee schedule PDF/CSV into a CPT rate table
- Add manual override editor for the top 50 CPTs
- Store rates per payer per contract
Deliverable: Populated rate table for 3 sample contracts
Compute expected minus paid per claim line
- Match claim lines to contracted rates by CPT plus modifier
- Compute variance and flag above threshold
- Render the variance table in the dashboard
Deliverable: Upload to underpayment table working end to end
Ship the sellable artifact
- Generate per-payer PDF letter, line-item CSV and evidence appendix
- Wire $499 Stripe checkout with token access
- Deploy the production build
Deliverable: MVP live with downloadable packet and working checkout
Reduce manual cleanup per audit
- Test 5 additional EOB formats from public samples
- Add per-field confidence flags and a manual correction cell
- Add a processing audit log
Deliverable: 10 EOBs processed with no manual cleanup required
Make the offer sendable
- Produce one redacted sample dispute packet
- Record a 90-second Loom walkthrough
- Publish a one-page offer page
Deliverable: Sample packet, Loom and offer page live
Week 2 · Launch & first sale · 7 days
200 named prospects
- Build 200-row LinkedIn list of billing managers at 2-10 provider specialty practices
- Tag rows by specialty and state
Deliverable: 200-row outreach list with message fields
First 40 sends
- Send 40 connection requests with the free-audit offer
- Log every reply in a single tracker
Deliverable: 40 sends and first replies logged
Deliver found-money numbers
- Deliver 3 free 20-claim audits
- Ask each recipient for a 10-minute reaction call
Deliverable: 3 sample packets delivered to prospects
Volume plus follow-up
- Send 40 more messages
- Follow up on all day-9 non-replies
Deliverable: 80 cumulative sends
Convert the warmest audit recipient
- Send the $499 checkout link to the warmest audit recipient
- Offer a 15-minute packet walkthrough
- Repeat for the next two warmest threads
Deliverable: First paid customer targeted
Test the group channel
- Post the free-audit offer in 3 medical billing Facebook groups
- Attach the redacted packet as proof
Deliverable: 3 group posts live with replies tracked
Remove the top objection heard
- Rewrite the offer page from objections heard
- Tighten the message template
Deliverable: Updated offer page and message template
Week 3 · Repeat & systematize · 7 days
Message variant test
- Send 40 messages
- Test Loom-first against text-first messaging
- Follow up with week-2 audits
Deliverable: 120 cumulative sends and variant data
Faster delivery per prospect
- Run audits against a 15-minute turnaround checklist
- Build a per-prospect found-money summary page
Deliverable: 5 more free audits delivered
Put a price in front of every warm thread
- Send follow-up 2 to all open threads
- Send checkout links to anyone who saw a packet
Deliverable: All warm threads carry a $499 ask
Offline conversations
- Call or visit 10 specialty practices within driving distance
- Ask for the billing lead by name
Deliverable: 10 local conversations and 2 audit commitments
Raise value per buyer
- Offer a second pack to buyers with over 1,000 lines
- Add a two-pack checkout at the same unit price
Deliverable: Repeat-pack offer live
Borrow the buyer's network
- Ask each buyer for one peer practice introduction
- Send that peer a free audit
Deliverable: Referral requests sent
Cut what does not convert
- Review close rate by specialty and message variant
- Drop the slowest-converting segment from outreach
Deliverable: Updated ICP note and messaging
Week 4 · Scale what works · 7 days
Concentrate on the best segment
- Send 60 messages to the best-converting segment
- Follow up on every thread from days 15-21
Deliverable: 180 cumulative sends
One-click packet assembly
- Automate packet assembly into a single job
- Add packet email delivery via Resend
Deliverable: Turnaround under 10 minutes per packet
Answer the security question before it is asked
- Publish anonymized variance examples from consenting users
- Add FAQ on PHI handling and retention
Deliverable: Trust page live
Refill the pipeline
- Deliver 6 free audits
- Ask for paid conversion on each one
Deliverable: 6 more packets delivered
Remove price risk
- Add refund language: no found money, no charge
- Document the 1,000-line scope line explicitly
Deliverable: Objection script and refund policy live
Resolve every open quote
- Call every warm prospect
- Close or close out outstanding quotes
Deliverable: All open quotes resolved
Keep regulated surface small
- Delete customer files after packet delivery per retention policy
- Confirm BAA coverage on storage and model providers
Deliverable: Retention and BAA checklist signed off
Days 29-30 · Consolidate · 2 days
Last uncontacted wave
- Send the final audit offer to 60 uncontacted prospects
- Post the closing group offer
Deliverable: Last outreach wave sent
Money before December 16, 13:00 UTC
- Collect all outstanding payments
- Send remaining packets and receipts
- Reconcile Stripe against TrustMRR
Deliverable: Final revenue collected and reconciled
Revenue and profit scenarios
How the numbers are defined
Hackathon profit — Collected revenue − refunds − chargebacks − payment processing fees − paid marketing spend.
Business profit — Hackathon profit − operating costs (hosting, domains, tooling, AI/API delivery costs).
The contest ranks entrants on hackathon profit only. Hosting, tooling and AI/API delivery costs are not deducted by the organisers, but each sale still has to cover them — so we show both. Definitions verified on 2026-11-15 against the official rules.
Stated assumptions
- Projection only, not verified revenue, customers or traction (all scenarios).
- Price is fixed at $499 one-time per audit pack in every scenario (all scenarios).
- Payment processing assumed at Stripe 2.9% + $0.30, about 3.2% effective (all scenarios).
- Conservative: 4 paying customers from roughly 180 outreach messages and 8 free audits; no refunds, no chargebacks.
- Base: 12 paying customers, one refund, no chargebacks, no paid marketing; free audits do most of the selling.
- Upside: 30 paying customers, two refunds, one chargeback, $300 of paid testing spend on sponsored posts.
- Operating cost covers S3 storage with retention, GPU/API extraction tokens, Vercel and Postgres; assumed $15-$30 per audited customer.
- No EHR integration, no sales team, no paid contracts, so no cost lines beyond those listed.
- Ad spend is paid marketing only; organic LinkedIn and group outreach is $0.
- The concept's 200-customer figure is treated as a ceiling, not a plan; the base case is 12.
Conservative
4 customers- Refunds
- $0
- Chargebacks
- $0
- Payment fees
- $64
- Paid marketing
- $0
- Hackathon profit
- $1,932
- Operating costs
- $85
- Business profit
- $1,847
Base case
12 customers- Refunds
- $499
- Chargebacks
- $0
- Payment fees
- $192
- Paid marketing
- $0
- Hackathon profit
- $5,297
- Operating costs
- $150
- Business profit
- $5,147
Optimistic
30 customers- Refunds
- $998
- Chargebacks
- $499
- Payment fees
- $479
- Paid marketing
- $300
- Hackathon profit
- $12,694
- Operating costs
- $320
- Business profit
- $12,374
Model your own numbers
Runs in your browser · nothing is saved- Collected revenue
- $5,988
- Payment fees (3.9%)
- $234
- Paid marketing
- $0
- Hackathon profit
- $5,754
- Business profit
- $5,604
Hypothetical arithmetic, not a forecast. The fee rate is a flat 3.9% assumption, refunds and chargebacks are left at zero here, and none of this is tracked data.
All three scenarios are projections derived from stated assumptions. They are not verified revenue, customer counts or traction.
Main risks and validation experiments
Risk 1PHI handling and BAA requirements block or delay selling to real practices.
Redaction preview, short retention with deletion after packet delivery, BAA-eligible storage, and a de-identified sample mode that works without real PHI.
Confirm in writing which storage and model providers offer BAA terms at solo-founder pricing before day 5; run the first free audits on de-identified samples if needed.
If no BAA-eligible path exists under roughly $100 per month by day 5, restrict the product to de-identified CSV remittances only and re-test demand.
Risk 2Extraction accuracy collapses on unfamiliar EOB layouts and produces a wrong variance number.
Per-field confidence flags, a manual correction cell, and a founder review of every free audit before it ships.
Run 10 publicly available sample EOB formats through extraction by day 6 and count clean claim lines.
If clean line accuracy stays below 80% after day 6, ship CSV-only input and drop PDF parsing from the pitch.
Risk 3A new solo-founder brand is not trusted with billing documents, so nobody accepts the free audit.
Free audit with no full-record upload, redacted sample packet, and a no-found-money-no-charge refund line.
Send 40 LinkedIn messages by day 9 and count accepted audits; target at least 2 by day 10.
If 120 messages by day 15 produce fewer than 3 accepted audits, the message or the buyer is wrong and the segment must be narrowed.
Risk 4Underpayments found are too small to justify $499, so pipelines stall at the price.
Target specialty practices with high per-claim values and audit 1,000 lines instead of 20 on the paid pack.
Measure found money per 1,000 claim lines across the first five audits.
If the median found amount is under $500 across five audits, narrow to higher-value specialties rather than cutting the price.
Risk 5Refunds or chargebacks exceed projections and erase hackathon profit.
Explicit scope line of 1,000 claim lines, one revision included, refund only when no underpayment is found.
Track delivery-to-refund on every pack; review at day 21.
If refunds exceed 20% of sales by day 21, stop selling until the accuracy gap is fixed.
Hour-by-hour: the first 24 hours
Hours 0-2
- Create and verify Stripe account, connect it to TrustMRR
- Confirm no customers or payments exist before launch
- Register domain and create the Vercel project
Output: Connected payment account and a deployed app shell URL
Hours 2-5
- Scaffold Next.js upload page and claim-line JSON schema
- Write the 20-claim audit spec with acceptance criteria
- Open a Postgres instance and job table
Output: Repo with upload UI, schema file and audit spec
Hours 5-8
- Run GPT-4o extraction on 3 sample EOBs
- Print extracted claim lines for manual inspection
- Note the first three failure modes
Output: Extraction spike output and a failure-mode note
Hours 8-10
- Draft the LinkedIn message template and the free-audit offer
- Build the first 40 rows of the prospect list
Output: Message template plus a 40-row outreach list ready for day 9
Full 30-day launch report
Read the full generated report
Underpay Scout: 30-Day Launch Report
This report is a plan, not a record. Nothing here describes existing customers, revenue, partnerships or verified results. All financial figures are projections built on stated assumptions.
1. Product opportunity and positioning
Underpay Scout audits claims that were already paid. The billing office posts the EOB, accepts the allowed amount, and moves on; nobody re-reads the contract fee schedule for line 14 of a 60-line remittance. That is where sub-$100 underpayments live, and because each one is small, no one builds a case around them. Underpay Scout builds the case automatically: remittance lines in, contract variance out, and a packet the practice can actually send to the payer.
The buyer is one person: the billing manager or practice administrator at a 1-10 provider US specialty practice who posts EOBs and holds the payer contracts. The trigger is concrete: month-end close, a contract renegotiation, or a new manager told to find revenue without new headcount. Those buyers answer LinkedIn messages; they do not run procurement.
JEV scored this concept 53.3/100 (band C, confidence 0.62). The strongest criteria are speed to first revenue (7.3 normalized) and reachable customers (6.33); the weakest are operational simplicity (1.91) and market demand (4.18). That profile dictates everything below: sell fast with a credible free audit, and keep the regulated surface deliberately small rather than adding capability.
Positioning against the three adjacent options is clean. A billing service takes a percentage of collections and wants the whole revenue cycle; Underpay Scout charges one flat fee and wants one upload. A clearinghouse submits claims and has no opinion about contracted rates. Denial-appeal tools chase claims that were rejected, not claims that were paid short. Underpay Scout sits between posting and write-off, and its output is a dispute packet, not an appeal letter.
What makes it sellable in 30 days is the demo format. The founder runs 20 claims for free and returns a number. If the number is zero, the prospect lost ten minutes. If it is $700, the $499 packet sells itself. No integration, no pilot, no production security review.
2. Smallest sellable MVP
One upload flow and one output. The buyer drags remittance PDFs and a fee schedule into the browser, sees a redaction preview, and receives a variance table plus a downloadable packet. Everything else is out: no EHR connection, no clearinghouse, no payer portal automation, no denial appeals, no patient billing, no RCM dashboard, no 24/7 promise. Manual rate override for the top CPTs stays in, because fee schedules are messy and the variance has to be defensible. A visible processing audit log stays in, because trust, not features, is the first blocker.
3. Technical implementation plan
Next.js on Vercel, Postgres for jobs and accounts, S3 for short-lived documents, pdf.js for preview, GPT-4o for remittance extraction, fee-schedule parsing and letter drafting, Stripe for the single $499 checkout, Resend for delivery. Four build days: extraction spike, rate ingestion, variance engine, packet plus checkout.
The known break point is remittance layout variance. The mitigations are a small per-payer template library, per-field confidence flags, and a manual correction cell in the variance table so a bad parse is fixed in seconds instead of failing the job. Dollar figures are always computed by the deterministic variance engine, never by the model.
4. Pricing and monetization
One product: Underpayment Audit Pack, $499 one-time, up to 1,000 claim lines, per-payer packet, variance report, one email revision. High enough to signal a real audit, low enough to clear without procurement, and one recovered line often exceeds it. Delivery cost per pack is a few dollars of extraction tokens and storage, so margin holds at any volume. Repeat revenue comes from additional packs at the same price; no subscription is introduced during the contest.
5. First-customer acquisition
The channel is LinkedIn direct outreach to billing managers and practice administrators at 2-10 provider specialty practices: 40 messages per day, a free 20-claim audit as the offer, a redacted packet and a found-money number as the proof. Supporting surfaces, not replacements: three medical billing Facebook groups and ten local practices within driving distance. Connection request, message, four-day follow-up, then a price. Cold email is not used and no ads are bought.
6. 30-day execution shape
Days 1-5 build and ship, with Stripe connected before any code so no payment ever lands outside the contest window. Days 6-7 harden accuracy and produce the sample packet, Loom and offer page. Days 8-14 build the 200-row list, send 120 messages, deliver five free audits and target the first paid close on day 12. Days 15-21 test a Loom-first message, add local calls, repeat packs and referral asks. Days 22-28 automate packet assembly, publish a trust page, resolve every open quote, and tighten retention and BAA paperwork. Days 29-30 are a final outreach wave, collection and Stripe-to-TrustMRR reconciliation before December 16, 13:00 UTC.
7. Revenue and profit scenarios
Three projections at the fixed $499 price: conservative 4 customers ($1,996 gross), base 12 ($5,988), upside 30 ($14,970). Processing is modeled at roughly 3.2% (Stripe 2.9% + $0.30). Base carries one refund ($499); upside carries two refunds ($998), one chargeback ($499) and $300 of paid testing spend. Operating cost runs $85 to $320 covering storage, extraction tokens and hosting. Hackathon profit subtracts refunds, chargebacks, processing and ad spend only; business profit additionally subtracts operating cost. These are hypotheses, not commitments.
8. Main risks and validation experiments
PHI handling is the top risk and is tested in week one by confirming BAA-eligible storage and model terms before a single real EOB is processed. Extraction accuracy is tested by running ten public EOB formats and counting clean lines; below 80% after day 6, input drops to CSV only. Trust is tested by 40 messages on day 9 and a target of two accepted audits by day 10. Value size is tested by found money per 1,000 lines across five audits; a median under $500 means the specialty is wrong, not the price. Refund risk is capped with an explicit scope line and a no-found-money-no-charge rule.
9. First 24 hours
Connect Stripe to TrustMRR, deploy the app shell, write the audit spec, then run a three-EOB extraction spike and start the first 40 prospect rows. The day-one artifact is a live URL, a connected payment account, extracted claim lines in a database, and a message ready to send on day 9.
This brief was generated from public data about established startups, then narrowed and scored for a solo 30-day launch. Revenue and profit figures are scenario models, not verified results, guarantees or evidence of existing customers. Validate demand yourself before you spend a month on it — and check the contest rules for what counts toward your score.
Related ideas
All ideas →Renewal Diff Brief
For independent employee-benefits brokers, Renewal Diff Brief turns a client's current plan PDF and carrier renewal packet into a cited employer change brief and employee explainer deck in minutes.
ShortPay Sifter
ShortPay Sifter ingests a small practice's ERA/835 files or claims export and returns a per-payer list of claims that paid below that practice's own historical allowed rate, packaged as a ready-to-send underpayment dispute packet.
ClaimLeak Teardown
Turns a de-identified claims CSV into a white-label Revenue Leakage Audit PDF that independent medical billing companies use to win small-clinic clients.