AI agent for journalists
Source Verification for Breaking Claims Agent
Show the reporter which claims have independent confirmation and which do not, as the story develops.
What it does
On a breaking story, claims come in fast from officials, witnesses, social posts and wires, and the reporter has little time to see which have two independent sources. This agent keeps a live list of the claims in the story. For each claim, it records where it came from and searches for corroboration: official statements, documents, wire services, local outlets, images and public data. It rates each claim as confirmed by two independent sources, single source, conflicting or unverified, and flags those that are unverified. It rechecks as new information arrives and shows what changed. It never publishes. The journalist approves what runs. Edge case: two outlets repeating the same statement count as one source, not two.
How it works
Follow the arrows from top to bottom. The orange dashed arrow is the loop: when a check fails, the agent goes back and tries again.
Read the steps as a list
- Claim added or new information arrives
- List the claim and record its origin
- Search for corroborating sources
- Check whether the sources are independent of each other
- Do two independent sources support the claim?If not: search further for primary documents or direct sources, then mark single source. Back to step 3.
- Rate the claim and note any conflicts
- Flag unverified claims for the reporter
- Journalist decides what to run and how to attribute itThe agent waits here for your OK.
- Watch for updates
- Has new information changed a rating?If not: recheck the sources and update the rating. Back to step 3.
- Claim verification table
How it decides
A claim is confirmed only when two independent sources, not copies of each other, support it. A conflicting source moves the claim to conflicting until resolved.
- Count a source as independent only if it did not copy another
- Rate conflicting sources as conflicting until resolved
- Flag any claim with a single source for attribution
- Recheck all claims when a major update arrives
Make it yours
Every agent is a starting point. You choose these settings for your own situation.
- Independence rules
- Source types to search
- Rating labels
- Update interval
- Newsroom verification policy
What keeps you in control
It always asks you first
- What is published
- How each claim is attributed
Hard limits
- Never publish or contact sources
- Show the source behind every rating
It stops when
- Done: story filed with ratings reviewed
- Stop: deadline passes and the reporter takes over
Set it up
We guide you through the set-up, step by step
Members get the full set-up guide for this agent. No technical skills needed: you copy, paste and upload.
- One set of instructions to paste into your AI, with the clicks for ChatGPT, Claude, Microsoft 365 Copilot, Gemini and Grok
- The agent then walks you through connecting your own data, one source at a time
- A downloadable copy with the flow chart, the rules and the full guide