How we use AI

You will always know

Any message, reply or post that Eclipse sends on its own is marked as coming from an AI system operating for Eclipse. The system never signs as a person and never presents itself as a human employee.

What the system does on its own

• Scores organisations on public registry data and visible business signals. It does not score people and does not use personal characteristics.
• Drafts outreach, replies, content and reports from verified evidence. Every claim traces to a source record.
• Adjusts campaign settings each week inside limits a person has set in advance.

What a person always decides

• Whether anything is sent to a new audience, on a new channel, or with more budget.
• Whether any public-interest content is published.
• Any reply to a person who has shown interest, before it goes.
• Pricing, contracts and anything with a legal effect.

Lines the system cannot cross

Eclipse will not target or optimise on vulnerability, will not generate synthetic media without a provenance record, and will not change these rules itself. They are held outside the system's reach and changed only by the people who run Eclipse, with a record of who changed what.

Models and data

Drafting uses hosted open-weight language models selected by region. No personal data is sent to providers outside the UK and EEA for any workload that touches it. Read the privacy notice for what we hold and your rights.