Back to case studies
Healthcare Ops

From denial rework to claims triage inside the EHR

A 27-person specialty clinic group was losing hours to avoidable claim denials. We shipped triage inside the existing EHR to flag risky submissions, assemble supporting evidence, and draft appeals for qualified staff to review.

Timeline:8 weeks · EHR boundary to production handover
Delivery:Production handover
Healthcare operations workspace with organized records

Project snapshot

The small-team operating context

The delivery and value signals below are anonymized and rounded so the operating scale is visible without exposing the client.

Client context

Specialty clinic group · 27 employees · 2 locations · ~900 claims per month

Delivery team

2 engineers + 1 clinical product lead with the billing manager and compliance lead

Sponsor

Revenue Cycle Manager

First-pass denials

One-third fewer

Reported in eight weeks

Project-reported

Claim volume

~900 / month

Small-clinic operating baseline

Directional model

Submission authority

100% human-owned

Qualified staff approve every submission

Project-reported

Production signal

  • High-risk claims were flagged before submission instead of waiting for payer denial.
  • Appeal drafts arrived with supporting records and rationale for staff review.
  • Denial outcomes fed the evaluation set for rules, extraction, and risk classification.

What we learned

  • Build the denial taxonomy with billers and coders who see the edge cases daily.
  • Keep clinical and coding evidence traceable to the source record.
  • Treat compliance, access, and audit controls as part of the workflow from week one.
Delivered: EHR-native triage workflow · Reviewer queue and appeal drafting · Audit controls, runbook, and denial feedback loop

Actual process teardown

How the operating process changed

The important change is not a model in isolation. It is the full operating path around the model: inputs, orchestration, controls, review, and the system of record.

The challenge

Five revenue-cycle staff across two specialty clinics were working through roughly 900 claims a month. Problems often surfaced only after payer review, forcing staff to search the EHR, payer response, coding fields, and clinical documentation before rebuilding an appeal case by case.

The production response

We connected structured claim and clinical context through a FHIR boundary, added deterministic pre-submission checks and AI-assisted risk review, and drafted evidence-backed appeals inside the EHR. Qualified staff retained final submission authority.

Before · manual baseline

  1. 01

    Submit the claim

    Clinic staff

    Claims with incomplete, inconsistent, or risky fields entered the payer workflow before staff could catch them.

  2. 02

    Wait for the denial

    Payer + clinic

    Problems surfaced after payer review instead of at the point where the claim could still be corrected.

  3. 03

    Investigate the reason

    Revenue cycle team

    Staff searched the EHR, payer response, coding fields, and supporting clinical records manually.

  4. 04

    Draft an appeal

    Revenue cycle team

    Appeal language and evidence were assembled case by case, with the risk of missing a supporting record.

After · AI-assisted production

  1. 01

    Read the EHR context

    System

    The workflow receives structured claim, encounter, coding, and relevant clinical context through the existing FHIR boundary.

  2. 02

    Flag risk before submission

    AI + rules

    Rules and AI-assisted review surface missing, inconsistent, or high-risk details with source context.

  3. 03

    Draft the appeal evidence

    AI

    The system assembles a reviewable draft with supporting records, rationale, and unresolved evidence gaps.

  4. 04

    Review and submit

    Human reviewer

    A qualified staff member validates the clinical and coding evidence and owns the final submission.

Phase-by-phase delivery

What was delivered, in order

Each phase turns the process teardown into a concrete production deliverable, with a clear value created before the next layer is added.

  1. 01Delivery phase

    Map denial and rework

    Find which missing, inconsistent, or risky fields create avoidable first-pass denials.

    Delivered

    • Claim-to-denial process map
    • Payer response and denial taxonomy
    • Baseline for review effort and submission outcomes

    Value created

    The workflow focuses on preventing rework at the point where staff can still change the claim.

  2. 02Delivery phase

    Connect the EHR boundary

    Bring the right structured clinical and claim context into the existing staff workflow.

    Delivered

    • FHIR normalization boundary
    • Claim, encounter, and coding context model
    • Role and audit controls for sensitive records

    Value created

    Staff do not need to copy clinical context into a separate tool to triage a claim.

  3. 03Delivery phase

    Build pre-submission triage

    Use rules and AI-assisted checks to surface risk before the payer sees the claim.

    Delivered

    • Missing-field and consistency checks
    • Risk classification and prioritization
    • Evidence-backed appeal draft generation

    Value created

    The team can fix cleanable issues early and reserve deeper appeal work for the cases that need it.

  4. 04Delivery phase

    Review, submit, learn

    Keep qualified staff accountable while using outcomes to improve the checks and drafts.

    Delivered

    • Reviewer queue inside the EHR
    • Final submission and audit update
    • Denial feedback loop for rules and model evaluation

    Value created

    Quality improves through a closed operational loop rather than an isolated prediction score.

How to read this teardown: the delivery sequence combines the existing public-safe project narrative with a practical view of the production work around it. It does not add client names, confidential implementation details, or unverified performance claims.

Measured and modeled impact

Separate the outcome from the value hypothesis

Project-reported outcomes are shown alongside directional models. Public benchmark figures are kept in their own evidence section below.

First-pass denial rate

Project-reported

One-third fewer first-pass denials

Before

Denial-heavy submission process

After

Pre-submission risk triage in the EHR

Interpretation boundary

Anonymized and rounded project-reported outcome supplied for this case study; independently unverified in the public view.

Value model
Directional model

Recovered reimbursement and reduced rework

For a small specialty clinic group, the return is recovered reimbursement and fewer hours spent rebuilding claims after the payer has already rejected them.

Model

Recovered claims × net reimbursement value + avoided rework hours × loaded billing cost − incremental review cost

Assumptions to validate

  • The ~15-hour weekly capacity estimate is validated through time sampling.
  • Claim status and denial reason data are available for a baseline.
  • Recovered reimbursement is separated from timing acceleration.
  • Clinical and billing reviewers retain final submission authority.

ROI scorecard

Value across time, money, resource, output, and error

Each teardown shows the practical ROI lens for the transformation. Existing project signals stay labeled; financial impact remains a transparent model until baseline volumes, costs, and quality data are verified.

Time
Project-reported

Not published

The public project summary reports fewer first-pass denials, not a measured staff-time reduction.

Money
Directional model

Recovered reimbursement

Directional model: recovered claims × net reimbursement value, less incremental review and submission cost.

Resource
Directional model

~15 hrs/week potential

For five revenue-cycle staff, earlier risk review can return roughly two working days per week from denial investigation to clean submission work.

Output
Project-reported

More clean first passes

Fewer avoidable denials create more claims that can move through the workflow without rework.

Error
Project-reported

One-third fewer denials

The existing project report says first-pass denial rate fell by a third in eight weeks.

ROI should be finalized against the client baseline: loaded labor cost, transaction volume, unit economics, throughput, and measured error or denial rates. The directional model is a value pathway—not an invented client result.

System blueprint

The production system around the AI

The tool list shows what each component is responsible for across the application, data, AI, integration, orchestration, and governance layers.

Tools used

Application

TypeScript

EHR workflow, reviewer queue, and appeal drafting UI

Interoperability

FHIR

Structured claim, encounter, coding, and clinical-context boundary

Data + audit

Postgres

Claim, risk, review, appeal, and audit state

Infrastructure

Azure

Production runtime, identity, secrets, and security boundary

Integration

EHR adapter

Event triggers, source-record links, and approved workflow updates

Governance

Rules + validation engine

Required fields, coding consistency, risk thresholds, and reviewer routing

Orchestration

  1. 1EHR event trigger
  2. 2FHIR normalization
  3. 3Rule and model checks
  4. 4Risk classification
  5. 5Appeal draft generation
  6. 6Qualified reviewer approval
  7. 7Submission and audit update

In-house data and boundary

The system is grounded in the client’s existing operational data rather than an unbounded external corpus.

EHR patient and encounter contextClaim and payer response dataClinical documentation and coding fieldsPrior denial and appeal history with role-based access controls

Human in the loop

  1. 1Review every high-risk claim before submission.
  2. 2Validate clinical and coding evidence against the source record.
  3. 3Reject drafts that cite incomplete, unsupported, or irrelevant records.
  4. 4Own final appeal and payer communication; the system does not diagnose or submit autonomously.
AI capabilities
Claim risk classificationStructured field validationEvidence retrievalAppeal drafting

Public benchmark evidence

What the wider market is showing

These sources support the transformation pattern. They are not client-specific outcomes and should not be read as such.

Public benchmark

CAQH — 2024 Index Report

CAQH documents substantial cost and savings opportunities when manual healthcare administrative transactions move to electronic workflows.

Why it matters here

Provides an adjacent industry benchmark for the cost of manual claim-administration work; it is not a measured result for this case.

Public benchmark

CMS — Optimizing Care Delivery Framework

CMS frames administrative burden reduction as a way to return time to clinicians and improve timely care delivery.

Why it matters here

Supports measuring returned staff time and workflow burden alongside reimbursement outcomes.

External research provides context and benchmarks. It does not verify the anonymized client, implementation, or project-specific results described above.

Names, logos, and specific numbers are redacted in this public view. Full references are available under NDA on request.

Build the next one

Have a workflow that needs a closer look?

Start with the friction you already see. We will help you determine what is worth solving and how to move it safely into production.

Assess my workflow Talk to us