Platform

One console for policy, risk, and coverage.

Four pillars that security and IT teams use to roll out protection and stay ahead of adoption — plus the employee-facing transparency the law and common decency both ask for.

Command Center

Posture, rollout status, and quick paths into what needs attention. The setup checklist walks the compliant path: publish policy, deliver notice, record attestation, then deploy.

  • Org posture at a glance
  • Setup checklist with the compliance order built in
  • Extension download & packaging

Policy

Publish human-readable standards. Every decision anchors to an active version — with history, so you can show exactly what the rule said on the day it was enforced.

  • Version history with effective dates
  • Decisions cite the policy section
  • Notice templates per state, built in

Risk

Event timeline and investigation views with outcomes expressed in your policy language — category-level by default, verbatim text strictly role-restricted.

  • Event timeline and investigation views
  • Category-only storage by default
  • Role-scoped access, every read logged

Coverage

See who is protected, who still needs the extension, and what to deploy next. Rollout progress without guessing.

  • Per-employee extension status
  • MDM and Enterprise policy rollouts
  • Agent coverage for AI outside the browser

Trust built into the platform

My AI Usage — transparency for every employee

Every monitored employee gets a page showing everything collected about them, without asking anyone. They can see their practice areas, export their data, and read coaching guidance in plain language. No scores, no rankings — coaching language only.

Security posture without the theater

AES-256 at rest, TLS 1.3 in transit, row-level security isolating each organization. No SOC 2 report yet — we say so, and offer verifiable narrowness in its place.

We train our own models

Coaching models are trained in-house. Customer prompts are never sent anywhere for training, and verbatim prompt text is never a training input. Read the model-training disclosure for the full picture.