Agentic Trust IndexATI Screen v1.0 Beta

ATI Verify Coming Soon

Trust claims should become testable claims.

ATI Verify will evaluate the technical evidence supporting an ATI Screen assessment, moving each claim from what a team says to what its configuration, records and tests show.

Future evidence

What ATI Verify is designed to examine

  • IAM / RBAC

    Identity and permission policies granted to the system.

  • MCP configurations

    Servers, tools and scopes the system can connect to.

  • API specifications

    The interfaces through which the system acts.

  • OPA / Rego

    Policy-as-code evaluated before consequential actions.

  • Tool definitions

    What each tool can do, and with what parameters.

  • Architecture

    Diagrams showing enforcement points and paths around them.

  • Audit logs

    Records that reconstruct an action end to end.

  • Approval workflows

    How approval is bound to the specific action.

Where Verify sits on the evidence ladder

  1. 01

    Assertion

    Someone states that the control exists.

  2. 02

    Documented

    A design, policy or procedure describes the control.

  3. 03

    Implemented

    Current configurations or records show the control is in place.

  4. 04· Reported

    Tested

    Dated test results show the control operates, including attempts to bypass it.

  5. 05

    Independently Verified

    A qualified party outside the delivery team examined the evidence.

Verify targets the step from self-reported claims to tested, evidence-supported claims.

Pilot

Join the ATI Verify Pilot

We are looking for a small number of teams with agentic systems in pilot or production to help shape evidence requirements. Request early access below.

Do not include secrets, credentials, or confidential architecture details.

Describe the system at a high level. We will agree how to handle sensitive material before any is shared.

Opens your email app with the request pre-filled.