AnnouncementTurnkeeper is working toward an open standard for sharing safety intelligence across platforms.

Pilot program

Join the design-partner waitlist.

A focused pilot runs 30 to 45 days. Setup requires one target workflow, example scenarios or policy docs, review-owner mapping, and one integration path. Success is measured by fewer risky AI actions, faster review cycles, and cleaner audit evidence.

Design partner waitlistOpen

Outcomes and raw conversations remain yours

How it starts

From one surface to a shared operating record.

We scope design-partner work around one harm trajectory, its review path, and the evidence your team needs to retain.

  1. Tell us about your surface

    Share the harm pattern, detector or platform signal, and specialist review path you want to connect first.

  2. Scope one trajectory

    We align on one evolving harm pattern, the privacy boundary, and the evidence standard your team can challenge.

  3. Map the review path

    Together we define the bounded event ontology, adapter surface, and explainable case workspace for authorized review.

  4. Evaluate together

    We review calibration results and operating boundaries before any expansion.

What to expect

Clear boundaries before broader deployment.

We describe the engagement by its operating boundaries—not by a future-state checklist.

One workflow, bounded scope

Design-partner work starts with one harm trajectory—not a platform-wide rollout or surveillance layer.

Metadata and provenance only

Turnkeeper organizes bounded signals, provenance, and review evidence. Raw conversations stay in your environment.

Explicit authorization

Shared signals never become unreviewed enforcement. Production effects stay customer-owned.

Elsewhere

Explore related surfaces.

Read the product boundary, research, and Field Notes before you commit.