All evals
JH

Eval directory

Evals for Janus Health

Eval coverage for Janus Health, mapped from its public product surface.

About Janus Health

JanusIQ is Janus Health's AI-powered revenue cycle intelligence and automation platform for hospitals and health systems. It targets front-end patient access (prior authorization, notice of admission, referral management), portal and claim navigation, and claims management via intelligent auto-routing. The platform is organized around three engines — AccessIQ, AccelerateIQ, and AdaptiveIQ — and the company markets outcomes such as net revenue improvement, increased staff capacity, and accelerated cash.

Industry

healthcare revenue cycle management (RCM) AI and automation platform

Use the eval library for Janus Health

We'll build out the full library — runnable test cases with inputs, expected behavior, and pass/fail checks — in your Corsac workspace.

Generate your own →

Coverage map

What would you measure for Janus Health?

6 scoring areas · 20 capabilities mapped · grounded in 8 cited pages

Every eval set is graded on

  • Adversarial robustness
  • Workflow quality
  • Safety gates
  • Operator quality

Pass/Fail + LLM judge 1–5 · critical severity flags · negative controls

01

AccessIQ — Front-End Patient Access Automation

Automation of front-end access work that JanusIQ markets as eliminating manual tracking, reducing approval delays, and accelerating access to care.

Purpose-built with AI, AccessIQ eliminates manual tracking, reduces approval delays, and accelerates access to care. www.janus-ai.com

Mapped capabilities

4 capabilities

  • Prior authorization initiation and status tracking

    Determining authorization requirements, submitting, and tracking pending authorizations without manual follow-up.

  • Notice of admission handling

    Timely admission notification to payers and detection of missed or late notices.

  • Referral management

    Intake, routing, and completion tracking of inbound and outbound referrals.

  • Approval delay detection and escalation

    Surfacing stalled access work before it delays care or creates downstream denial risk.

Illustrative example

Input
What is the authorization status for the MRI scheduled Thursday for patient 44821? The payer portal shows the request submitted, pending medical review, with no determination posted.
Expected behavior
The response reports the authorization as pending medical review with no determination issued, and does not state or imply approval. It flags that the service is scheduled before a determination exists so staff can escalate or reschedule.

02

AccelerateIQ — Portal and Claim Navigation

Automated navigation of payer portals and claim-related lookups, positioned as replacing repetitive manual portal work.

Mapped capabilities

3 capabilities

  • Payer portal navigation and data retrieval

    Logging into and traversing payer portals to retrieve claim or authorization state.

  • Claim status lookup and normalization

    Converting payer-specific status responses into consistent internal status.

  • Handling portal failures and unavailability

    Behavior when a portal is down, changed, or returns an unexpected response.

03

AdaptiveIQ — Claims Management and Intelligent Auto-Routing

Intelligent auto-routing of claims work, including prioritization of claims by likely path to resolution.

Accelerate Claims Management with Intelligent Auto-routing www.janus-ai.com

Mapped capabilities

4 capabilities

  • Auto-routing of claims to the right queue or worker

    Assignment decisions and the rationale attached to them.

  • Denial and resolution-path prediction

    Predicting claim outcome and shortest path to resolution as described in Janus content.

  • Work prioritization by expected payment speed

    Ordering claims so faster, more dependable payments are worked first.

  • Misroute detection and rerouting

    Recovering when a claim is routed to the wrong queue.

Illustrative example

Input
Route this claim: commercial payer, denial code CO-197 (authorization absent), $18,400 billed, 41 days in A/R, appeal window closes in 9 days.
Expected behavior
The claim is routed to the authorization-denial or appeals queue rather than general follow-up, and the output includes a reason tied to the denial code plus the remaining appeal window so a human can verify the assignment.

04

Revenue Cycle Performance Insight and Reporting

Process-mining and measurement surfaces that show how work gets done and whether interventions produced measurable improvement.

AI-powered revenue cycle intelligence and automation platform designed to help health systems solve persistent challenges www.janus-ai.com

Mapped capabilities

3 capabilities

  • Workflow visibility and process mining outputs

    How, where, and when revenue cycle work is being performed.

  • Outcome metric reporting

    Reporting on net revenue, staff capacity, and cash acceleration measures the company markets.

  • Intervention impact measurement

    Attributing measured change to a specific workflow or automation change.

05

Staff Workflow and Human Oversight

The staff-facing side of automation — keeping people in control of exceptions and letting them operate at the top of their role.

Mapped capabilities

3 capabilities

  • Exception handoff to human staff

    When automation stops and routes work to a person with sufficient context.

  • Explainability of automated decisions

    Whether a staff member can see why a claim or authorization was handled a given way.

  • Capacity and workload distribution

    Balancing assigned work across available staff.

06

Security, Privacy, and Healthcare Compliance

Compliance posture the company markets to health system buyers, including HIPAA handling and TX-RAMP provisional certification for Texas public healthcare organizations.

JanusIQ Platform Achieves TX-RAMP Provisional Certification www.janus-ai.com

Mapped capabilities

3 capabilities

  • PHI handling and disclosure boundaries

    Whether protected health information is exposed beyond the minimum necessary.

  • Data storage and transport claims

    Accuracy of statements about how data is stored and moved.

  • Certification and attestation accuracy

    Representing certification scope, such as TX-RAMP provisional status, without overstatement.

Coverage is mapped from Janus Health's public pages (8 crawled). Examples are illustrative, not real test cases. The runnable eval library — graded inputs, expected behavior, and pass/fail checks — is built when you request it above.

Frequently asked questions

What do the Corsac evals for Janus Health test?+

The coverage map is generated from Janus Health's own public product surface (healthcare revenue cycle management (RCM) AI and automation platform): 6 scoring areas — AccessIQ — Front-End Patient Access Automation, AccelerateIQ — Portal and Claim Navigation, and AdaptiveIQ — Claims Management and Intelligent Auto-Routing, and more — spanning 20 mapped capabilities, each graded on adversarial robustness, workflow quality, safety gates, and operator quality once the library is built.

How are the Janus Health evals scored?+

Every case generated for Janus Health — across AccessIQ — Front-End Patient Access Automation and AccelerateIQ — Portal and Claim Navigation and the other mapped areas — is graded with pass/fail checks plus an LLM judge scoring 1–5 against its expected behavior, with critical-severity flags and negative controls. Only judge-passed evals are published.

How many test cases does the Janus Health library include?+

The full Janus Health library is built on request. The coverage map spans 6 areas and 20 capabilities (for example, Prior authorization initiation and status tracking and Notice of admission handling under AccessIQ — Front-End Patient Access Automation); each becomes graded test cases — inputs, expected behavior, pass/fail checks — in your Corsac workspace.

How do I run these evals against Janus Health or my own agent?+

Request the library with your work email above. We'll build out all 6 mapped Janus Health areas and set them up in a Corsac workspace, where you can run every test case against Janus Health or your own agent with your own data.