AI Disclosure

Effective Date: April 20, 2026 Last Updated: April 20, 2026

Aumata is a hybrid marketing service. A human growth strategist designs your strategy; AI agents execute the daily work. We believe in being explicit about where AI is and isn’t in the loop, for you, for your customers, and for our own compliance with the EU AI Act transparency obligations that apply from August 2, 2026.

1. Where we use AI

We use large language models (primarily Anthropic’s Claude family) for the following purposes:

Content generation: SEO/AEO articles, blog posts, landing page copy, social media posts, email sequences, and ad creative drafts produced by our content agents.
Visibility and competitive research: summarizing SERP data, competitor activity, and keyword clusters for your dedicated strategist.
Internal operations: code, documentation, and internal tooling that we use to run the business.

In every case, AI output that reaches public or customer-facing channels is produced or reviewed in collaboration with a human strategist before publication. No AI agent acts as the final approver for content that carries your brand.

2. Where we do not use AI

We do not use AI to make “significant decisions” about individuals (within the meaning of the CPPA’s Automated Decision-Making Technology rules): financial, employment, housing, healthcare, education, or essential-goods access decisions.
We do not use Customer or end-user personal data to train or fine-tune AI models. Our AI sub-processors (including Anthropic) are contractually prohibited from using data transmitted via API for model training.
We do not generate deepfakes, impersonations, or content designed to mislead about the source of a statement.

3. How we disclose AI-generated content

Articles auto-published by our SEO agent to customer sites are reviewed by the assigned strategist prior to merge. Customers may require all content to carry an “AI-assisted” label; we support that on request.
For content that falls within the EU AI Act’s “public-interest information” category (for example, news, political, or health reporting), customers are required by our Master Service Agreement to apply an AI disclosure label where their local law requires one.
Visitors to aumata.ai should assume that most long-form blog content on this site was drafted with AI assistance and reviewed by Nick Vossburg or a delegated editor before publication.

4. Your rights when AI is used on our Services

Ask for human review: any customer or prospect who receives content, a strategy audit, or a lead-qualification outcome influenced by our agents may request human review by emailing [email protected]. We will route the request to Nick or a senior strategist within five business days.
Opt out of agent-generated outreach: marketing emails sent with AI assistance carry standard unsubscribe links. Unsubscribing opts you out of both human and AI-generated outreach.
Correct mistakes: if you see inaccurate AI-generated content about you or your business on this Site, email [email protected]. We will correct or remove within a reasonable timeframe consistent with our correction obligations under applicable privacy law.

5. Known limitations

Large language models sometimes produce inaccurate, outdated, or hallucinated content. We mitigate this through multi-pass generation, rubric scoring, retrieval grounding in real data (DataForSEO, Ahrefs, Firecrawl), adversarial risk classification before publication, and human review. No amount of engineering eliminates the risk entirely. If you find an error, tell us; we’ll fix it.

6. Models we rely on

Claude Opus 4.6 — long-form content generation, complex reasoning
Claude Sonnet 4.6 — planning, routing, and moderate-complexity agent steps
Claude Haiku 4.5 — monitoring, scoring, lightweight classification

Models and providers may change. The current sub-processor list is at www.aumata.ai/legal/subprocessors.

7. Contact

Questions, corrections, or requests for human review:

Red Sovereign LLC d/b/a Aumata Email: [email protected]