Demo & live evidence

See the governance workflow run on a real book.

The walkthrough below shows a full onboarding end-to-end - manuscript intake through Canon Pack construction, governance copilot intervention, A/B preview, and accept-and-launch. Not a slide deck. Not a canned demo.

§ 01   The workflow, run on a real book

End-to-end onboarding, recorded.

Use SPACE or arrow keys to advance · Open full screen ↑

§ 02   What CP-Lite does

Editorial governance in controls your editors already think in.

  1. 1

    Title-level interpretive profile

    Each book gets a structured editorial object: thesis, voice, what the AI may do, what it must not do, how to handle ambiguity. Your editors set this through a guided workflow - not free-form prompts.

  2. 2

    Anticipatory + corrective governance

    Before launch, your editors specify the misreadings most likely to appear. After launch, our governance copilot watches for drift and proposes corrections in A/B preview. Nothing changes silently.

  3. 3

    Portable across platforms

    The governance layer travels with the title. Embed in your reader app. Expose through your own API. License to a partner platform. Same editorial voice everywhere your title meets a reader.

§ 03   When the AI drifts

You have a workflow — not a hope.

The single biggest publisher concern about AI-mediated reading is what happens when an output is wrong, off-tone, or off-brand. CPLite's answer is not "we'll do our best." It is an operational governance loop:

  1. Detect

    The governance copilot watches every companion response against the title's red-line rules.

  2. Propose

    When drift appears, the copilot proposes a specific bounded correction - warmth, directness, scope, hard boundary, or vocabulary.

  3. Preview

    The proposed change is shown in A/B form: the same question answered under current and proposed configurations.

  4. Decide

    Your editor accepts, rejects, or revises. Logged. Reversible.

  5. Repeat

    Per title, as drift evolves and reader behavior reveals new edges.

This loop is what publishers are licensing - not a static configuration, but an inspectable, revisable governance practice.

Ready to engage?

Three engagement paths, from self-serve evaluation to per-title onboarding. Pricing tiers and what stays with you.

See pricing - engagement → Read the methodology