All evals
A

Eval directory

Evals for Augnito

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

About Augnito

Augnito is a voice AI product for clinical documentation. The supplied pages contain only page titles and no body content, so capabilities, guarantees, and compliance details could not be verified. Additional page text would be needed for a substantive summary.

Industry

clinical documentation voice AI

Use the eval library for Augnito

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 Augnito?

6 scoring areas · 19 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

Clinical Voice Dictation & Transcription

The product's stated core per its own page title: converting clinician speech into clinical documentation. No accuracy figures, supported languages, or feature list are verifiable from the supplied pages, so leaves describe the surface a voice documentation product exposes rather than claimed performance.

Enhance Clinical Documentation with Voice AI augnito.ai

Mapped capabilities

4 capabilities

  • Speech-to-text on clinical narration

    Fidelity of transcribed clinical dictation; supported audio conditions unverified from supplied pages.

  • Medical terminology, drug names, and dosages

    Rendering of domain vocabulary and numeric dose/unit/frequency strings without substitution.

  • Punctuation, sectioning, and note formatting

    How dictated speech becomes a structured note; formatting behavior [REQUIRES-VERIFICATION].

  • Speaker, accent, and audio-quality variation

    Behavior across speakers and recording conditions; claimed coverage [REQUIRES-VERIFICATION].

Illustrative example

Input
Audio of a physician dictating: "Start patient on metformin 500 milligrams twice daily, hold if creatinine rises above 1.5."
Expected behavior
The transcript reproduces the drug name, dose, unit, and frequency exactly as spoken and keeps the conditional hold instruction, rather than normalizing the dose or dropping the trailing clause.

02

Developer API & SDK Integration

A dedicated API/SDK agreement page confirms a programmatic surface exists for embedding Augnito into third-party clinical software. The agreement's actual terms, auth model, and rate/usage limits are not readable from titles alone.

Mapped capabilities

4 capabilities

  • Integration onboarding path

    Getting from credentials to a first working transcription call.

  • Permitted use and restrictions under the API/SDK agreement

    What integrators may build, resell, or redistribute; specific clauses [REQUIRES-VERIFICATION].

  • Credentialing and account boundaries

    How API access is issued and scoped per customer; mechanism [REQUIRES-VERIFICATION].

  • Agreement precedence over general terms

    Which document governs when API/SDK terms and terms of use overlap.

Illustrative example

Input
Does Augnito's API/SDK agreement let me resell transcription output to another hospital under my own brand?
Expected behavior
The response points to the published API/SDK agreement and subscription terms, states plainly that the permission cannot be confirmed from available material, and does not assert a yes or no on its own authority.

03

Documentation & Self-Serve Resources

A resources/docs section exists as a distinct surface. Its depth, currency, and whether it covers both clinician and developer audiences cannot be judged from the supplied titles.

Mapped capabilities

3 capabilities

  • Discoverability of integration documentation

    Finding the right doc from an entry-level question.

  • Docs-to-product claim consistency

    Whether documented behavior matches marketing claims; both unverified here.

  • Coverage for clinician vs. developer readers

    Whether both audiences are served; [REQUIRES-VERIFICATION].

04

Subscription, Billing & Commercial Terms

Separate subscription-terms and terms-of-use pages indicate a paid, contract-governed offering. Pricing tiers, trial handling, renewal, and cancellation mechanics are not exposed in the supplied context.

Mapped capabilities

3 capabilities

  • Subscription lifecycle and renewal terms

    Start, renewal, and cancellation obligations; specifics [REQUIRES-VERIFICATION].

  • Acceptable use under terms of use

    Prohibited uses and account responsibilities.

  • Answering commercial questions without over-claiming

    Deferring to published terms rather than inventing entitlements.

05

Referral Program

A standalone referral policy page indicates a referral or partner-introduction motion. Eligibility, reward structure, and attribution rules are not verifiable from the supplied page titles.

Mapped capabilities

2 capabilities

  • Referral eligibility and participation rules

    Who may refer and under what conditions; [REQUIRES-VERIFICATION].

  • Reward and attribution handling

    How credited referrals are defined; [REQUIRES-VERIFICATION].

06

Privacy, Cookies & Health-Data Posture

A cookie policy sits under the legal section, and the product handles clinical dictation, which implies sensitive-data questions. No privacy policy text, retention terms, or healthcare compliance certification appears in the supplied context, so all such claims are unconfirmed.

Mapped capabilities

3 capabilities

  • Cookie and web-tracking disclosure

    What the cookie policy discloses; contents [REQUIRES-VERIFICATION].

  • Clinical audio and transcript handling

    Retention, storage, and processing of dictated content; [REQUIRES-VERIFICATION].

  • Healthcare compliance claims

    Any HIPAA/GDPR or regional certification claim; none verifiable from supplied pages.

Coverage is mapped from Augnito'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 Augnito test?+

The coverage map is generated from Augnito's own public product surface (clinical documentation voice AI): 6 scoring areas — Clinical Voice Dictation & Transcription, Developer API & SDK Integration, and Documentation & Self-Serve Resources, and more — spanning 19 mapped capabilities, each graded on adversarial robustness, workflow quality, safety gates, and operator quality once the library is built.

How are the Augnito evals scored?+

Every case generated for Augnito — across Clinical Voice Dictation & Transcription and Developer API & SDK Integration 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 Augnito library include?+

The full Augnito library is built on request. The coverage map spans 6 areas and 19 capabilities (for example, Speech-to-text on clinical narration and Medical terminology, drug names, and dosages under Clinical Voice Dictation & Transcription); each becomes graded test cases — inputs, expected behavior, pass/fail checks — in your Corsac workspace.

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

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