Clean ClAImsFirst Pass

CombineHealth vs Trase

Two RCM Automation Platforms vendors, side by side. Facts from public sources; judgments are ours.

At a glance

Derived from public facts · a rough scale, not a ranking

CombineHealthTrase
Pricing model

Per-transaction / per-chart · Priced per chart or claim worked

Enterprise contract (custom) · No upfront fee, billed on efficiency gains

Speed to go live

Sidecar agents on existing EHR, weeks

Prebuilt agent library deploys in weeks

Automation model

Autonomous agents · Named agents with human escalation

Autonomous agents · Governed agent OS for regulated work

Built for

Mid-size groups, Billing companies

Enterprise systems

Security posture

SOC 2 Type II, HIPAA

SOC 2 Type I, HIPAA

Company maturity

4 yrs (est. 2022)

Not disclosed

Financial backing

Pre-seed (Y Combinator W23)

$117.5M ($10.5M pre-seed plus $107M seed) · Seed

Named customers

5 named

2 named

Published results

Specific numbers public

Specific numbers public

Documented integrations

3 listed

None documented

Third-party validation

None found

None found

Bottom line

  • Pick CombineHealth if you want an early-stage, aggressively priced AI agent workforce across coding, billing, and denials and can accept startup vendor risk.
  • Pick Trase if you are an academic or large health system that wants governed AI agents across back-office workflows and prefers paying from measured efficiency gains.

CombineHealth

Named AI agents that work the revenue cycle end to end

Founded
2022
HQ
San Francisco, CA
Stage
Pre-seed (Y Combinator W23)
Raised
$500K

What it does

  • Amy codes encounters with rationale in 2-4 minutes
  • Mark preps claim-ready charges and posts payments
  • Adam works denials, payer portals, and IVR calls
  • Rachel drafts and files appeals
  • Taylor surfaces revenue cycle analytics and leakage
  • Human-in-the-loop escalation with explainable decisions

Where it's strong

  • Covers front, mid, and back office with one agent framework instead of point tools.
  • Publishes concrete accuracy claims (97% coding accuracy, 20% denial reduction on 10,000+ claims).
  • Small YC-backed team moves fast and prices aggressively against legacy RCM vendors.

What buyers should weigh

  • Only $500K disclosed funding and a small team, so vendor viability is a real diligence item.
  • Customer list skews to mid-size groups and ER networks, not large health systems.
  • Agent autonomy claims need validation in your specialty and payer mix before scaling.

Named customers

Union Health · McFarland Clinic · SignatureCare ER · Brault · Medcor

Integrations

EHR and PM systemsPayer portalsClearinghouses
Full CombineHealth profile →

Trase

Governed AI agents for healthcare back-office work

Founded
n/a
HQ
McLean, VA
Stage
Seed
Raised
$117.5M ($10.5M pre-seed plus $107M seed)

What it does

  • AI agents for inbound fax triage and routing
  • Patient access workflow automation
  • Revenue cycle task automation
  • HIPAA and SOC 2 compliant agent platform
  • Governance and oversight controls for agent actions

Where it's strong

  • Live health system deployment at Duke, where fax triage ran 7.1 times faster than manual processing and saved about 1,395 staff hours a month.
  • The $107M seed led by ARCH Venture Partners gives it unusual capital depth for a company at this stage.
  • Pricing tied to measured efficiency gains lowers the risk of paying for software that does not perform.

What buyers should weigh

  • The company only came out of stealth in November 2025, so the reference base beyond Duke Health is thin.
  • The platform spans healthcare, defense, and other regulated industries, so healthcare-specific product depth may lag single-focus rivals.
  • Published results center on fax triage; patient access and RCM claims have less public evidence.

Named customers

Duke Health · U.S. Navy

Full Trase profile →

Compare against the rest of RCM Automation Platforms

Deciding between these two?

First Pass tracks RCM Automation Platforms every week: funding, launches, and what changed since this page was written.