Skip to content
LAUNCH FILM — LIVEGREENLIGHT · PATENT-PENDINGRUNTIME AUTHORIZATION FOR AI AGENTSMCP CONNECTORS — IN DESIGNISO/IEC 27001 — ROADMAPSOC 2 TYPE II — ROADMAPISO/IEC 42001 — ROADMAPMAS FEAT — DESIGN-ALIGNEDDETERMINISTIC · EXPLAINABLE · TAMPER-EVIDENTAI ACTS · CRELIS DECIDESLAUNCH FILM — LIVEGREENLIGHT · PATENT-PENDINGRUNTIME AUTHORIZATION FOR AI AGENTSMCP CONNECTORS — IN DESIGNISO/IEC 27001 — ROADMAPSOC 2 TYPE II — ROADMAPISO/IEC 42001 — ROADMAPMAS FEAT — DESIGN-ALIGNEDDETERMINISTIC · EXPLAINABLE · TAMPER-EVIDENTAI ACTS · CRELIS DECIDES
Crelis.ai
Future extension

An illustrative concept, not part of the current product. Crelis today decides allow, require human approval, escalate, or block; when policy requires human approval, the reviewer is one your organisation designates.

Human reviewer marketplace · concept

Human judgment, on demand

An optional, future extension of runtime authorization: a pool of authorized reviewers an organisation could draw on when policy returns require human approval and no in-house reviewer is available.

Apply for Pilot Access

Illustrative reviewer pool

What an authorized reviewer pool could look like

Filter the illustrative pool by sector. Each profile shows a role, expertise, availability, and clearance — role-only profiles, no real individuals, no volumes or review-time figures. Illustrative data.

Illustrative data · future extension

Illustrative reviewer pool · role-only profiles, no real individuals

How it could work

Judgment as infrastructure

Not a help desk: a routed, accountable seat for human decision-making inside the authorization loop — if and when this extension is built.

Reviewer onboarding

Reviewers would pass identity, credential, and domain checks before being authorized to review a single task.

Availability-aware routing

Routing would consider who is qualified and available, so an action waiting on human approval reaches a reviewer who can act on it.

Accountable decisions

Each reviewer decision would join the same tamper-evident record as the agent action it approved or declined — reviewer role, context, and outcome together.

Your pool or a shared one

An organisation could draw on its own authorized reviewers, or on a shared pool for sectors where qualified reviewers are scarce.

Where humans plug in

A defined seat in the authorization loop

Authorized reviewers enter at the require-human-approval stage with full task context, and their sign-off is sealed into the same audit record as the AI's work. This step is live today for reviewers your organisation designates; the shared pool above is the future extension.

Human approval path

Illustrative data

  1. 01

    AI action

    Agent proposes a consequential action

  2. 02

    Deterministic policy

    Inputs evaluated; the policy that applies fires

  3. 03

    Require human approval

    Outcome when policy calls for sign-off

  4. 04

    Authorized reviewer decides

    Full context; sign-off joins the record

  5. 05

    Visa issued or withheld

    Signed execution visa; your system executes

  6. 06

    Evidence recorded

    Tamper-evident record written

Apply for Pilot Access