AI Data Processing

• summarise publicly available articles, podcasts, posts, and reports;
• score or rank content by relevance to a client;
• recommend journalists, creators, companies, or opportunities;
• draft outreach, campaign copy, newsletters, and reports;
• analyse client-uploaded documents for the relevant workspace;
• generate briefs, insights, and recommendations;
• generate social content, image previews, and media prompts.

Depending on the feature, we may send:

• prompts and instructions;
• short article excerpts or public-source text;
• links, titles, bylines, metadata, and summaries;
• client-provided content needed for the requested output;
• organisation profile details;
• style guides, writing samples, and brand context;
• professional contact context;
• draft outreach or campaign content;
• social reports, media prompts, and generated media previews;
• user instructions and edits.

We aim to send only the information reasonably necessary for the requested feature.

We do not use Client Content to train foundation models. We do not opt in to share Client Content with AI providers for model improvement, and we do not permit Client Content to be used for provider foundation model training under our commercial/API arrangements.

Current product features may use OpenAI, Anthropic, and Google AI services. OpenAI is used for summaries, recommendations, drafting, analysis, and selected web search workflows. Anthropic is used for selected article generation, refinement, voice-of-customer, thought-leadership, and style-analysis workflows. Google AI services, including Gemini, Imagen, and Vertex AI, are used for selected social content, image preview, and media generation workflows.

Provider-specific standard retention may apply for abuse monitoring, security, service operation, debugging, or legal compliance unless eligible zero data retention or equivalent controls are approved and enabled for the relevant provider, project, and endpoint. We do not publicly claim zero data retention unless it has been confirmed for the relevant provider, project, and endpoint.

If we add or replace a material AI provider for Client Personal Data, we will update this page, update our Subprocessors page, and provide subprocessor notice where required by our Data Processing Addendum or client agreement.

Some workflows use AI-assisted source discovery. In the current configuration, Big Ears AI uses OpenAI web search for selected research tasks and Perplexity for limited domain and publication research. Perplexity receives domain names, publication names, and related public-source queries rather than Client Content or connected mailbox data.

AI-generated output may be inaccurate, incomplete, biased, or out of date. Users are responsible for reviewing outputs before relying on them, publishing them, or sending them externally.

Outreach should be reviewed by a human before it is sent. Clients remain responsible for ensuring their outreach complies with applicable law and professional standards.

AI outputs and workspace learning are scoped to the relevant client workspace. Big Ears AI does not use one client's confidential materials, connected mailbox data, or campaign strategy to generate outputs for another client.

Questions about AI data processing can be sent to james@bigears.ai.