AI Disclosure

How AI is used here.

VoidAi uses AI agents (running on the Anthropic Claude API) to draft content, classify inbound mail, build proposals, and run ops infrastructure. This page explains what that means for you, what data flows where, and what humans review before anything reaches you.

What's AI-drafted on this site

• Build Log posts. Drafted by the content-publisher agent. Reviewed by the founder before publication. Marked with an AI-disclosure footer on every post.
• Service descriptions. The offer-builder agent maintains the offer copy in 03-Revenue/models/. Reviewed by the founder before site rebuild.
• Email replies you receive. Drafted by the customer-ops agent. Always reviewed and sent by the founder. No agent ever auto-sends.
• Agent roster page. Pulled at build time from each agent's worker.md charter. Static at the moment the site rebuilds.

What data flows to Anthropic

When you submit the contact form, the body of your message (problem-worth-solving + what-tried + budget-range + timeline) passes through the Anthropic Claude API for classification and kickoff-email drafting. Per Anthropic's API terms:

• — Your data is not used to train Anthropic models.
• — It's encrypted in transit and at rest.
• — It's retained per Anthropic's data-handling policy at anthropic.com/legal/privacy.

If you require us to redact your data before it enters the Anthropic API (HIPAA-style requirements, sensitive client lists, etc.), tell us in the contact-form message and we'll configure redaction in the customer-ops + design-queue workflows before kickoff.

What humans review

Every customer-facing artifact is human-reviewed:

• — Email replies: drafted by agent, sent by Jeff.
• — SOWs: drafted by offer-builder, signed by Jeff after legal review.
• — Build Log posts: drafted by content-publisher, reviewed by brand-steward agent and then Jeff.
• — Site copy changes: drafted by content-publisher or Jeff directly, deployed by web-deployer.

The 6-row Guardian red-line list explicitly forbids email.send, payment.execute, and share.public for every agent. The capability gate is checked before any tool call; the audit log preserves the receipt.

CA AB 730 (AI in advertising)

California requires disclosure when AI is used to generate content for advertising. This page is that disclosure. Every Build Log post also carries an inline AI-disclosure footer.

Questions

Email jeff@voidco.com directly — that mailbox bypasses every agent. We're happy to walk you through the audit log, the agent charters, the Guardian gate, or any other piece of the operation.

Last updated: 2026-04-25