Project Cecurus — AI-powered Claims Integrity

Catch billing errors before they leave your hospital.

A HIPAA-safe, on-premise AI platform for claims integrity — for hospitals and specialty practices. Real-time detection of coding errors, duplicate claims, and prior-authorization mismatches. PHI never leaves your environment.

Join the Waitlist See how it works →

The Scale of the Problem

U.S. healthcare loses $68 billion a year to billing errors — and the smallest clinics pay the biggest price.

Here's where the loss actually lands.

Total System — Annual Loss
$68B

Lost annually across U.S. healthcare to billing errors, claim denials, and improper payments. Most of it preventable — catchable if someone reads the claim before it leaves the building.

Council for Affordable Quality Healthcare (CAQH), 2023
Average U.S. Hospital
$11M

Lost per year to denied claims and the cost of appealing them. Up 23% year over year.

Premier Inc., 2024
Specialty Clinic — Annual
$1M

Lost per year at a typical 10-provider specialty practice — enough to threaten independent survival.

MGMA benchmarking

Up to 90% of denied claims are preventable — catchable if someone reads the claim before it leaves the building.

— American Hospital Association, Cost of Caring Report

Behind the numbers: doctors stuck on paperwork instead of patients. Small clinics closing because they can't keep up with insurance. Surprise bills that never should have been sent. Cecurus catches the mistakes before they become someone's problem.

How It Works

Three steps. Zero PHI exposure.

Cecurus sits between your billing system and the clearinghouse — analyzing every outbound claim inside your environment, before it ever reaches a payer.

EHR / RCM System
837 Claim Generated
◈ CECURUS — On-Premise
Clearinghouse
Payer
01
Deploy On-Premise

A lightweight agent installs inside your existing environment. It monitors your outbound 837 file path — no EHR API credentials, no workflow disruption, no cloud data transfer. Setup takes days, not months.

02
Intercept & Analyze

Every claim is analyzed in memory the moment it's generated. Our AI + rule engine checks for coding errors, missing modifiers, prior-auth mismatches, medical necessity gaps, and duplicate claims — using only the minimum fields required.

03
Review & Report

Flagged claims surface in a local investigation interface for your billing and compliance teams — with plain-language explanations of every flag. Aggregate metrics appear in a cloud dashboard with no PHI, ever.

A local investigation queue, running inside your environment.

When Cecurus flags a claim, this is what your billing and compliance team review — a plain-language explanation of what the rule engine caught, the estimated risk, and one-click actions. PHI never leaves the local browser.

Cecurus
cecurus.local / claims queue
D. OKOYE
DO
Workspace
Queue7
Rules
Audit Log
Reports
Settings

Claims queue

7 FLAGGED · 41,208 SCANNED THIS CYCLE
SCANNING
Claim ID Category Severity Est. risk Status Detected
h7f2a…9c1b PRIOR AUTH HIGH $4,820 Pending review 2026-08-10 09:14
RULE FIRED: rule.PRIOR_AUTH_MISMATCH CONFIDENCE: 0.94

Prior authorization on record does not cover CPT 27447 billed for this date of service.

Recommended action: verify auth or resubmit with correct auth number.

Assign for review Open in EHR Dismiss
PHI RENDERED LOCALLY · NOT TRANSMITTED
b3c81…4e77 DUPLICATE MED $3,210 Pending review 2026-08-10 08:52
a91d4…2f30 CODING HIGH $6,140 Pending review 2026-08-10 08:31
e44b0…8a15 PRIOR AUTH LOW $1,450 Resolved 2026-08-09 17:06
c05e9…1d62 CODING MED $2,975 Pending review 2026-08-09 15:44
f28c7…60ba DUPLICATE LOW $1,880 Dismissed 2026-08-09 11:20
d61f3…7c48 PRIOR AUTH MED $3,640 Pending review 2026-08-09 09:58
Real-time queue · One row shown expanded for context

Detection Logic

What we catch.

Explainable AI + deterministic rules — designed to pass auditor review, tune to near-zero false positives, and evolve as our models learn from your claims.

Coding
Missing or Invalid CPT / Modifiers
Claims submitted with missing procedure codes, incorrect modifiers, or bundling errors that predictably trigger denial. Caught and flagged before submission.
Prior Auth
Prior Authorization Mismatches
Procedures submitted without required authorization, or with authorization that doesn't match the billed service or date range. One of the most preventable denial categories.
Necessity
Medical Necessity Gaps
Coded services without supporting documentation to justify medical necessity — the leading cause of post-payment recovery demands.
Duplicate
Duplicate & Overlapping Claims
Exact duplicates, line-level duplicates, inpatient/outpatient conflicts, and ED/inpatient overlaps — a category that alone accounts for meaningful annual leakage.
Eligibility
Eligibility & Coverage Mismatches
Claims filed against inactive coverage, non-covered procedures, or out-of-network scenarios. Catch before submission to preserve patient trust and clinic time.
cecurus — claim_screener.log
// 2026-07-15 14:23:41 UTC
// Claim intercepted — analyzing in memory

claim_ref: [HASHED]
patient_token: h7f2a...REDACTED
provider_token: p9c1b...REDACTED
dos_bucket: 2026-Q3
claim_type: 837P / Professional

// Running AI + rule engine...

rule.CPT_VALID: PASS
rule.PRIOR_AUTH: ⚠ FLAGGED
rule.MED_NECESSITY: PASS
rule.DUPLICATE: PASS
rule.ELIGIBILITY: PASS

// Alert generated (local only)
category: PRIOR_AUTH_MISMATCH
severity: HIGH
confidence: 0.94
est_risk_usd: $3,210

// No PHI written. No data exported.
✓ Routed to local investigation queue

Architecture

Privacy-first isn't a feature. It's the architecture.

Two strict trust boundaries. PHI never crosses the second one.

Inside Your Environment
On-Premise
EHR / RCM system
Claims data (PHI)
Detection engine
Investigation UI
Local audit trail
PHI never leaves
Aggregate
metadata only
One-way · no pull
Cloud Dashboard
Metadata only
Flagged counts
Risk totals
Rule metrics
Patient IDs
Dates of service
Free-text notes

Anchor Pilot

Real clinicians. Real workflows. Real proof.

Cecurus is preparing its anchor pilot with an orthopedic specialty group in Virginia — a high-volume, high-complexity coding environment where preventable denials cost the most and pilot results show up in weeks, not quarters.

Our physician advisor is embedded in the pilot site and in the design of the detection logic itself.

Pilot partner

Hampton Roads Orthopaedics Spine & Sports Medicine

Hampton Roads Orthopaedics
Spine & Sports Medicine

Williamsburg & Newport News, Virginia

Physician advisor

Dr. William Hill WH
Dr. William Hill, M.D., M.P.H.
Orthopedic surgeon at HROSM. Beryl Institute Patient Experience Scholar. Peer-reviewed researcher on physician–hospital integration.

The Team

Built by operators, clinicians, and researchers.

Four co-founders with backgrounds spanning healthcare finance, machine learning, data science, and operations — building the claims integrity platform we wish existed when we saw the problem up close.


Kelvin Baah KB
Co-founder

Kelvin Baah

The builder.

Owns Cecurus’s architecture, machine learning approach, and product design — the technical bet that lets detection happen on-premise without PHI leaving the environment. Prior startup co-founder. Currently pursuing a Ph.D. in Educational Policy, Planning, and Leadership.

Tracy Nyarkoah Obodai TO
Co-founder

Tracy Nyarkoah Obodai

The insider.

Works in accounting and finance at a U.S. hospital by day — every Cecurus product decision gets pressure-tested against how a revenue cycle team actually operates. Prior information security auditor (HIPAA, NIST, ISO). M.S. in Finance.

Kelvin Cecil Apenteng KA
Co-founder

Kelvin Cecil Apenteng

The bridge.

Sits between clinical partnerships and the data layer — calibrating detection against real claims and running the relationships that keep pilots moving. Previously in public health data analytics on Ghana’s COVID-19 self-testing program. Completing an M.S. in Data Analytics.

Who It's For

Built for the teams responsible when something goes wrong.

Three audiences. One platform. All on-premise.

Revenue Cycle Management

Catch errors before submission.

Reduce denied claims, rework costs, and clearinghouse fees. Get visibility into recurring billing patterns without adding headcount or centralizing patient data.

  • Pre-submission detection, not post-hoc auditing
  • Fewer denials, faster reimbursements
  • Recurring pattern visibility across claim types

Compliance & Audit

Demonstrate proactive compliance.

Maintain a local audit trail of every flagged event and reviewer decision. Generate standardized monthly summaries for regulators — without ever touching PHI in the cloud.

  • Explainable, deterministic rules for auditors
  • Local audit log — on-prem, never synchronized
  • Monthly compliance summary, no PHI included

Hospital IT & Security

Pass your security review.

Read-only directory access. No EHR API changes. No outbound PHI. Docker-based deployment that your security team can review, approve, and monitor.

  • Read-only, no workflow disruption
  • PHI-safe logging — no payload dumps in logs
  • Tenant isolation per hospital environment

Early Access

We're building this with hospitals, not for them.

Cecurus is in active development. We're working with a select group of hospitals and specialty practices to validate our architecture, refine detection logic, and shape the product roadmap.

  • Priority access to the beta — before public release
  • Direct input on detection rules and investigation workflow
  • Exclusive launch pricing locked in at signup
  • A seat at the table while the product is still being shaped

contact@projectcecurus.com

We don't share your information. No sales calls without your consent.

You're on the list.

We'll be in touch as we move toward early access. In the meantime, feel free to reach us directly at contact@projectcecurus.com