Use of AI

Decira uses AI to read and organize public records at a scale people couldn't manage alone. People review what funders rely on most, and anything they haven't reviewed is labeled.

Where we use AI

• Reading filings and the reports organizations publish
• Writing organizational summaries
• Sorting and screening news coverage
• Drafting material that an analyst then reviews

How it's labeled

• Anything an analyst hasn't reviewed is marked not analyst-reviewed.
• AI-written summaries of news coverage are marked as AI-generated, with the date they were written.
• Reviewed material shows the date of review.

What people decide

• Whether a possible sanctions match is a real match
• Whether news coverage is adverse to the organization
• What a review says
• Whether to fund. That decision is always the funder's.

Figures and ratings

Financial figures come from the filed returns and are calculated from them, not written by AI. Confidence levels and readiness tiers come from fixed rules applied to the filings and the media screen, not from AI. An analyst publishes them and can change them.

Your data

We don't use customers' data, or what customers enter in Decira, to train AI models, and the AI providers we use are not permitted to train on it either.

Mistakes

AI can be wrong. That's why unreviewed material is labeled and why every fact shows its source. If you find an error, see Corrections.

Last updated: October 15, 2026 · Questions: [email protected]