Option B — Intelligence Tier

The promise

Everything Option A does, plus AI: outline authoring, drafting, chart derivation, agentic generation, and optional web research with cited sources. AI calls are governed by a signed data-processing agreement, and every artifact is labelled contract-protected by the pipeline itself.

We will never call this tier “zero-egress” — because it isn’t. Content sent to the AI leaves your network through one governed channel, under contract. We think telling you that plainly is a feature.

How it works

1. Your team submits a topic, a brief, or existing content.
2. The DocMark box makes one governed HTTPS call to the Anthropic API — under your own Anthropic enterprise account (BYO key). Your security team clears Anthropic directly; by default, Anthropic does not use API data for training. Per-job spend caps apply.
3. Claude authors the outline or researches the topic — web research returns cited sources that ship in the deliverable.
4. Back on your box: deterministic branded render (pixels never from AI), then the local visual-QA gate. Pass and it’s delivered with its contract-protected label; fail and it’s withheld with findings.

The DPA story, as a feature

• BYO key (recommended): the AI relationship is yours, on your paper, cleared by your security team — we never see your content’s API traffic terms.
• Fail-closed attestation: if the DPA attestation isn’t in place, AI authoring refuses to run. There is no “oops, it called the API anyway.”
• Machine-enforced labelling: the pipeline writes the egress class on every artifact. A deliverable that touched the API can never carry a zero-egress label.

Who it’s for

Teams whose content is confidential, but whose security organisation can accept a contractual arrangement with an AI provider — and who want the productivity of AI authoring and cited research without giving up local rendering, local verification, and audit-grade provenance.