How to Build a Medical Billing and Revenue Cycle Platform in 2026

Medical billing is a $6B market in the US alone and the incumbents (Waystar, Change Healthcare, Availity, AdvancedMD, athenaCollector) either run on 2005-era UX or were acquired and stuck in integration purgatory. Independent practices, urgent care chains, specialty clinics, and mid-size hospital systems spend millions per year on RCM software that their staff openly hates. The rebuild opportunity is massive.

The business case has gotten sharper in 2026. Payer denial rates climbed to 12% industry-wide. ICD-11 is rolling out in phased adoption. Prior authorization volume doubled in three years. Surprise billing protections (No Surprises Act) created new workflow requirements. Every one of those changes is an opportunity for a modern platform that the incumbents are slow to adapt to.

The market splits roughly three ways. Independent physicians (250K practices) want Kareo-style simplicity. Mid-size groups (50 to 500 provider practices) want full RCM with analytics. Hospital systems want enterprise RCM with Epic and Cerner integration. Pick a segment and go deep. Trying to serve all three is how new entrants die. Related reading: how to build a healthcare app.

A medical billing platform automates the full revenue cycle. The workflow has eight distinct stages:

• Eligibility and benefits verification: Check patient coverage before the visit. Real-time 270/271 EDI transactions through a clearinghouse.
• Prior authorization: Submit auth requests for procedures that require it. X12 278 transactions plus fax for laggard payers.
• Charge capture and coding: Convert clinical notes into ICD-10 diagnoses and CPT procedure codes. Often AI-assisted in 2026.
• Claim scrubbing: Validate claims against payer-specific edits before submission. Thousands of rules per payer.
• Claim submission: EDI 837 transactions through a clearinghouse (Change Healthcare, Availity, Waystar, Trizetto).
• Remittance and payment posting: EDI 835 transactions come back with payment and denial details. Auto-post to patient accounts.
• Denial management: Review denied claims, identify root causes, generate appeals or corrections.
• Patient billing and collections: Statement generation, payment plans, online payment, collections workflow.

Every stage has data feeding into every other stage. Your platform is fundamentally a workflow engine over a healthcare data graph. Spend architecture time on that data model upfront.

Coding is the biggest source of revenue leakage in traditional practices. Under-coding leaves money on the table. Over-coding creates audit risk and potential fraud charges. AI-assisted coding is the new table stakes in 2026.

Code sets: ICD-10-CM for diagnoses (70,000+ codes), CPT for procedures (10,000+ codes), HCPCS for medications and supplies, ICD-10-PCS for inpatient procedures. Keep current copies licensed through AMA or Optum. Budget $15K to $30K per year for code set licensing.

AI-assisted coding: ingest clinical notes (often from an EHR via HL7 v2 or FHIR), run NLP models that extract diagnoses and procedures, map to codes, surface to the coder for review. Off-the-shelf options: 3M 360 Encompass, Nuance CAC, Dolbey Fusion CAC, plus newer entrants like Fathom, CodaMetrix, and Regard. Roll-your-own is viable with Med-PaLM 2, GPT-4o, or Claude combined with a medical coding specialist prompt.

Scrubbing: edit rules per payer. Medicare LCDs, Medicaid state-by-state rules, commercial payer policies. Vendors like Waystar and ZirMed provide these as a service; home-grown scrubbing is a 200 to 400 hour engineering project plus ongoing maintenance. We recommend starting with a commercial scrubber and layering your own logic on top.

Related: our HIPAA compliance cost guide covers adjacent infrastructure requirements.

You will not build your own clearinghouse connections. Every commercial medical billing platform integrates with one or more clearinghouses that handle the direct payer connections.

Clearinghouse options: Change Healthcare (largest, post-acquisition by Optum still has integration work in flight), Availity (payer-owned consortium, strong commercial coverage), Waystar (best developer experience in 2026), Trizetto (Cognizant-owned, decent), Office Ally (budget tier). Most practices use 1 to 3 clearinghouses depending on payer mix.

Integration: X12 EDI is the dominant format. 270/271 for eligibility, 278 for authorization, 837 for claims, 835 for remittance, 276/277 for claim status. Every clearinghouse wraps EDI in its own API (usually REST over JSON or SOAP) with varying levels of abstraction. Budget 200 to 400 hours for a production-grade integration with one clearinghouse, 80 to 160 per additional.

Eligibility: run real-time eligibility checks at appointment scheduling and again at check-in. Cache results for 24 hours. Parse 271 response for coverage, deductibles, copays, out-of-pocket maximums. Display in a clean UX for front desk staff.

Claim submission: nightly batch submission is typical. Pre-submit scrubbing, then package into 837 files, submit to clearinghouse, track acceptance at clearinghouse and at payer. Acknowledgments come back within hours (acceptance) to days (adjudication start).

Payment posting is where RCM staff spend most of their day, and where modern UX beats the incumbents.

835 parsing: ERA files come back with claim-level adjudications, line-item adjustments, payment amounts, reason codes, and denial codes. Parse these into structured records, match to submitted claims, and auto-post payments to patient accounts.

Reconciliation: compare expected payments (based on fee schedule and contract) to actual payments. Flag underpayments for review. Track contractual adjustments separately from denials. Integrate with bank deposit data to reconcile lockbox or ACH deposits to payer ERAs.

Secondary and tertiary billing: when primary pays, automatically generate secondary claims to Medicare, Medicaid, or supplemental insurance. Track coordination of benefits. Handle crossover claims (automatic Medicare-to-Medicaid or Medicare-to-supplemental).

Patient responsibility: after all payers process, calculate patient responsibility. Generate statements. Offer online payment via a portal (Stripe, Square, or healthcare-specific like Patientco or InstaMed). Support payment plans. Integrate with collections agency when needed.

A well-designed payment posting UX can cut staff time by 40 to 60% compared to incumbent platforms. This is a core differentiation lever for new entrants.

12 to 15% of claims get denied on first submission. Half of those can be successfully appealed. Denial management is where smart software pays for itself.

Denial taxonomy: build a classification layer that maps CARC (Claim Adjustment Reason Codes) and RARC (Remittance Advice Remark Codes) into actionable buckets. Authorization missing, coding error, eligibility issue, bundling, timely filing, duplicate claim, medical necessity.

Workflow per denial type: route authorization denials to a prior-auth team, coding denials to coding review, eligibility denials to front desk, medical necessity denials to provider appeals. Track SLAs and aging.

Automated appeals: for common denial patterns (Medicare NCDs, common coding mistakes, recurrent payer-specific denials), pre-build appeal templates. Pull relevant documentation (clinical notes, prior auth records), merge into appeal letter, submit via clearinghouse or fax. Even with AI, 30 to 50% of appeals still require human review.

Analytics: dashboard denial rates by payer, provider, CPT code, time of year. Identify systemic issues and fix them upstream. A 2% reduction in first-pass denial rate on a $10M practice is $200K in recovered revenue per year. Customers will pay premium for that analytics layer.

See our telemedicine app guide for adjacent healthcare integration patterns.

You are handling PHI (protected health information) at scale. HIPAA is non-negotiable and the violations are expensive ($100 to $50,000 per record depending on severity, capped at $1.5M per violation category per year, criminal penalties on top for willful neglect).

Privacy Rule compliance: minimum necessary access, patient rights (access, amendment, accounting of disclosures), BAAs (Business Associate Agreements) with every vendor that touches PHI, breach notification within 60 days.

Security Rule compliance: access controls (role-based, minimum privilege), audit logs (every PHI access logged for 6+ years), encryption at rest and in transit, secure disposal, workstation security, incident response plan, risk analysis and management program.

Infrastructure: HIPAA-eligible cloud services (AWS BAA, Azure BAA, GCP BAA), encrypted databases (AWS RDS with KMS, Azure SQL with TDE), PHI-redacted logging (do not log PHI in CloudWatch or similar), segregated production access.

Audit trails: every access, modification, submission, and decision logged with user, timestamp, action, and data accessed. Retained 6+ years. Queryable for audit response. Budget a dedicated audit log service (AWS CloudTrail plus custom PHI access logs in a separate database).

HITRUST CSF certification: not required by HIPAA, but hospital and large payer customers often require it. Budget $75K to $200K for the certification process plus ongoing assessor fees.

Medical billing software pricing typically follows one of three models. Pick based on your segment.

Per-provider SaaS: $200 to $800 per provider per month. Clean model for small practices. Easy to sell. Cap on upside.

Percentage of collections: 2 to 7% of collected revenue. Aligns incentives but creates customer scrutiny. Common for full RCM outsourcing.

Hybrid: Base SaaS plus per-claim fee plus percentage for denial work. Used by larger RCM platforms. Most flexible but harder to sell.

GTM channels: content marketing to practice administrators, partnerships with EHR vendors, direct sales to physician groups, channel partnerships with healthcare consultants, paid search (high-intent keywords, expensive but converts). Medical billing customers are highly loss-averse. Demo-to-close can be 60 to 180 days.

Team: founders should include at least one healthcare operator (former practice manager, biller, or MD). Budget 1 technical co-founder, 1 healthcare domain founder, 2 to 3 senior engineers, 1 designer, 1 compliance officer, 1 sales lead. Year-one burn $1.5M to $3M. Target $5M to $15M seed or Series A.

Revenue timeline: first paying practice by month 6 to 9. $500K ARR by month 18. $5M ARR by month 36 if you execute well. Medical billing is a slow-burn but highly defensible SaaS category with 95%+ net retention when you deliver.

Our SOC 2 for startups guide has more on compliance. If you want help scoping RCM software for your vertical or practice segment, book a free strategy call.