All evals
Tennr

Eval directory

Evals for Tennr

Eval coverage for Tennr, mapped from its public product surface.

About Tennr

Tennr is an agentic patient orchestration platform for healthcare providers that automates pre-visit patient processing — intake, documentation, payer criteria checks, and routing. It ingests patient data from any source (fax, portal, internal order), classifies and routes it, determines what documentation supports coverage and what is missing, and triages which cases need attention first. The company markets it as built on proprietary machine learning models developed by founders from Stanford AI research.

Industry

healthcare patient intake and referral orchestration AI

Use the eval library for Tennr

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 Tennr?

6 scoring areas · 24 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

Multi-Source Intake & Document Understanding

Ingesting patient data from fax, e-fax, email, e-prescribe, portal, and internal orders, then extracting structured order and patient information from messy real-world documents.

“Tennr automatically classifies and routes patient data from any source—fax, portal, internal order—into the right operational workflows.” www.tennr.com

Mapped capabilities

4 capabilities

  • Source-agnostic ingestion

    Accepting and normalizing orders arriving via fax, e-fax, email, e-prescribe, portal, and internally created orders.

  • Messy document extraction

    Reading handwriting, checkboxes, and low-quality scans without silently dropping or inventing field values.

  • Document classification

    Distinguishing referrals, orders, clinical notes, and non-clinical pages such as cover sheets.

  • Multi-order and multi-patient splitting

    Separating batched transmissions into the correct number of distinct patient orders.

Illustrative example

Input
A single inbound fax of six pages: a cover sheet, a referral for one patient, and a separate referral for a second patient with a handwritten date of birth.
Expected behavior
The system creates two distinct patient orders, one per referral, and does not treat the cover sheet as an order. The handwritten date of birth is either extracted or flagged as low-confidence rather than guessed.

02

Service-to-Payer Criteria Decisioning

Turning a library of payer criteria into operational determinations of what is covered, what documentation supports coverage, and what is missing — Tennr's stated path to slashing first-pass denials.

“Tennr is an agentic patient orchestration platform built for policy-grade decisioning and patient flow at scale.” www.tennr.com

Mapped capabilities

4 capabilities

  • Service-to-criteria mapping

    Matching the requested service to the applicable payer criteria for the patient's plan.

  • Documentation sufficiency

    Judging whether the assembled record supports coverage for the requested service.

  • Missing-evidence identification

    Naming the specific documentation gap rather than issuing a generic incomplete verdict.

  • Proof assembly

    Collecting and presenting the supporting documents behind a coverage determination.

Illustrative example

Input
An order for a service whose payer criteria require a qualifying test result, submitted with a physician note and demographics but no test result attached.
Expected behavior
The system determines the record does not yet support coverage and names the missing qualifying test result specifically. It does not assert that the service is approved or authorized on the record as submitted.

03

Intelligent Triage & Prioritization

Using clinical and operational context to decide which orders need attention now and which can flow, so staff effort lands where it matters most.

“Tennr combines agentic workflow automation, service-to-criteria mapping, intelligent triage, and quality-controlled autopilot” www.tennr.com

Mapped capabilities

4 capabilities

  • Urgency ranking

    Ordering the queue by clinical and operational signals rather than arrival time alone.

  • Attention routing

    Surfacing cases that require human decision-making versus those that can proceed automatically.

  • Decision guidance

    Explaining why a case was surfaced and what the reviewer should decide.

  • Stall detection

    Flagging orders that have stopped progressing between steps or queues.

04

Workflow Orchestration & Care-Setting Routing

Moving patients between consults, diagnostics, authorizations, and scheduling, and routing them to the right care setting across varying operational processes, staffing models, and escalation paths.

“Tennr automates outreach and follow-up across patients, payers, and providers with agentic communications” www.tennr.com

Mapped capabilities

4 capabilities

  • Care-setting routing

    Directing a patient or order to the appropriate setting and downstream workflow.

  • Configurable workflow fit

    Adapting routing to a given operation's process, staffing model, and escalation path.

  • Escalation handling

    Handing off to the correct person or queue when work cannot proceed automatically.

  • EHR integration behavior

    Writing and reading order state against the receiving provider's EHR.

05

Agentic Communications Coordination

Automated outreach and follow-up across patients, payers, and referring providers to collect missing information and share status without manual phone tag.

Mapped capabilities

4 capabilities

  • Missing-information outreach

    Requesting the specific missing item from the right party.

  • Referrer status visibility

    Keeping referring providers informed of where their patient stands.

  • Follow-up persistence

    Closing the loop on outstanding requests so no referral falls through the cracks.

  • Payer interaction handling

    Pursuing payer-side follow-up without stalling the order.

06

Autopilot Quality Control & Model Transparency

The quality-controlled autopilot boundary and the AI transparency Tennr markets: knowing when the system should act versus defer, and being able to explain how a determination was reached.

Mapped capabilities

4 capabilities

  • Autonomy boundaries

    Acting automatically only where confidence and policy allow, deferring otherwise.

  • Determination explainability

    Showing the criteria and source documents behind an operational decision.

  • Abstention on ambiguity

    Declining to assert a coverage or routing outcome that the record does not support.

  • Recovery from bad input

    Handling illegible, contradictory, or incomplete submissions without producing a confident wrong answer.

Coverage is mapped from Tennr'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 Tennr test?+

The coverage map is generated from Tennr's own public product surface (healthcare patient intake and referral orchestration AI): 6 scoring areas — Multi-Source Intake & Document Understanding, Service-to-Payer Criteria Decisioning, and Intelligent Triage & Prioritization, and more — spanning 24 mapped capabilities, each graded on adversarial robustness, workflow quality, safety gates, and operator quality once the library is built.

How are the Tennr evals scored?+

Every case generated for Tennr — across Multi-Source Intake & Document Understanding and Service-to-Payer Criteria Decisioning 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 Tennr library include?+

The full Tennr library is built on request. The coverage map spans 6 areas and 24 capabilities (for example, Source-agnostic ingestion and Messy document extraction under Multi-Source Intake & Document Understanding); each becomes graded test cases — inputs, expected behavior, pass/fail checks — in your Corsac workspace.

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

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