⌂ thingsthat.work Denial Fighter
A thingsthat.work build · by James Rodriguez
Revenue-integrity suite for healthcare denials

Stop losing revenue to denials —
diagnose, prevent, and stay ahead.

Denial Fighter is one system with three ways in. Find the denials already draining your revenue, catch the next one before the claim submits, and get ahead of the code changes that create tomorrow's denials. Adopt one module for immediate ROI — or all three for a closed loop.

Explore the three modules ↓ See the live engine ↗
Look back
"Where are we losing money?"
Diagnose systemic denial patterns in claims you've already filed.
Prevent now
"Will this claim get denied?"
Pre-check a prior auth against payer policy before it submits.
Stay ahead
"What's about to change?"
Track CMS code changes that will drive next quarter's denials.

Each module stands alone. Together they cover the full denial lifecycle.

Denials are systemic, point-of-care, and regulatory at once — no single tool covers all three. Denial Fighter maps one module to each.

🔍
Live prototype
Look back · Diagnose

Denial Pattern Intelligence Engine

Reads your historical claims and surfaces where denials cluster — by provider, diagnosis, payer, and time — then ranks the recoverable revenue. The analytical core, built and validated as an AMIA informatics project.

100,000 claims analyzed (11,476 patients)
4 / 4 denial patterns detected at 87–100% confidence
21 ranked, actionable recommendations
Open the live dashboard →
In development
Prevent now · Point of care

Denial Fighter Pre-Check

An AI agent that reads the relevant payer medical policy in real time, cross-references the clinical documentation, and returns a denial-risk score with the exact gaps to fix — before the prior auth is ever submitted.

FHIR-native — plugs into the EHR, no middleware
Gap detection against live medical-necessity criteria
Seconds to a fixable risk report
See how the pre-check works →
📬
Live · accepting subscribers
Stay ahead · Anticipate

Billing Code Watch

A monthly digest that monitors CMS publications and the Federal Register for billing-code changes, then explains each one's revenue impact and the action to take — so a code change never turns into a surprise denial.

CPT / HCPCS / ICD-10 + modifiers, auto-scanned
Impact + $ estimate + action on every change
Forward-looking — flagged before the effective date
Subscribe to the digest →

A closed loop around denied revenue.

The three modules aren't just adjacent — they feed each other. Findings from one sharpen the next.

🔍
Engine
Finds the patterns
Pre-Check
Stops the next one
📬
Code Watch
Warns before it changes

The Engine shows which payers and codes deny most often → those patterns become the highest-value rules for the Pre-Check → and Code Watch feeds both a live stream of policy changes, so the rules never go stale. The result is a system that gets smarter every month instead of fighting the same denials twice.

Start with one. Each stands alone — they just compound.

There's no all-or-nothing. Pick the module that matches your most urgent question today; add the others as they earn their place.

Adopt one

Immediate, standalone ROI

Run the Engine against last year's claims to recover money already left on the table. Or subscribe to Code Watch to stop losing revenue to code changes. Either delivers value on day one, alone.

Best for: "Prove it works on something real."
Adopt two

Diagnose + defend

Pair the Engine with Pre-Check: the patterns you find retrospectively become the rules that stop denials prospectively. Look-back and prevent-now, working off the same knowledge base.

Best for: "Turn our denial data into fewer denials."
Adopt all three

Closed-loop revenue integrity

The full system: diagnose systemic loss, prevent it at the point of care, and stay ahead of every regulatory change — a loop that self-updates as policy shifts. One vendor, one denial strategy.

Best for: "Own denials end to end."

A single progression from three prototypes to one platform.

The modules exist today as independent, working pieces. The plan connects them behind a shared knowledge base — without rebuilding what already runs.

Phase 1 · Now

Three modules, live and independent

Each piece already runs on its own and proves its own value.

  • Engine: validated on 100K synthetic claims, live dashboard + API
  • Code Watch: live, sending monthly digests, accepting subscribers
  • Pre-Check: policy-reading prototype + this front door
Phase 2 · Next

Shared payer-policy + code knowledge base

Connect the modules behind one source of truth so they stop duplicating work.

  • Code Watch changes flow into the policy layer the Pre-Check reads
  • Engine pattern findings become prioritized Pre-Check rules
  • One account + dashboard across all three modules
Phase 3 · Later

Closed-loop automation on real claims

Move from synthetic validation to a live pilot inside a real revenue-cycle workflow.

  • Engine re-scores denials continuously as new claims land
  • Pre-Check auto-updates when Code Watch detects a relevant change
  • Measured recovery rate reported back to the customer

Three steps from submission to clarity.

How the prevent-now module turns a prior auth into a risk report before the claim ever leaves the building.

01

Submit prior auth

Provider submits the request with supporting clinical documentation via FHIR — no new data entry.

02

AI reads the policy

The agent pulls the relevant payer medical policy and extracts medical-necessity criteria, documentation requirements, and exclusion rules.

03

Risk report

Returns a denial-risk score with the specific gaps, missing documents, and remediation steps — before the claim goes out.

One stack across the suite.

Python FastAPI LLM / RAG FHIR AI Agents Supabase CMS / Federal Register ingest Cron automation Email (Resend)

Want this running against your denials?

Denial Fighter is built and validated. If you own revenue cycle, medical policy, or coding at a payer or provider, let's pilot the module that fits your most urgent question.

Part of a workshop of healthcare & builder tools.

Denial Fighter is one of several things James has shipped. See the rest in the full gallery.

TradeDraftAI● Live
AI proposal builder for HVAC & electrical contractors — natural-language job in, polished proposal out.
tradedraftai.com →
ModelRegistry● Live
A registry and dashboard for tracking AI models across projects — versions, status, and health at a glance.
modelregistry.jamesrodriguez.dev →
The full gallery● Live
DocGuard AI, PolicyBot, Billing Code Watch, the Family Hub and more — every build in one place.
thingsthat.work →