AI agent for public relations specialists
Media Correction Evidence Agent
Correction requests that are justified and precise.
What it does
When news coverage looks wrong, correction requests are often poorly evidenced or aimed at opinions. When a team member flags an article, this agent first checks whether the disputed passage is a factual claim that can be proven wrong. Opinions are closed with a note. For factual claims, it compares the statement with dated, approved evidence. If the evidence clearly contradicts the claim, it drafts a narrow correction request for that sentence only. If the evidence is weak or only differs slightly, it explains why a request is not justified. It tracks the article for later changes. Media relations approves any contact with journalists. Edge case: an article calls the product overpriced, which is opinion.
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
- Article flagged
- Locate the disputed passage and its sources
- Is the disputed passage a factual claim?If not: close as opinion. Back to step 2.
- Compare with dated approved evidence
- Does evidence clearly contradict it?If not: explain insufficient support. Back to step 4.
- Draft a narrow correction request
- Media relations approves contactThe agent waits here for your OK.
- Documented correction case
How it decides
It checks whether the statement is falsifiable and supported by evidence before drafting.
- Opinions are never treated as factual errors.
- Evidence must clearly contradict, not just differ.
- Correction requests cover only the disputed sentence.
- No legal allegations in any request.
Make it yours
Every agent is a starting point. You choose these settings for your own situation.
- Approved evidence sources and how recent they must be
- Who can flag articles
- Tracking period after a request (default 30 days)
- Who approves contact with journalists
What keeps you in control
It always asks you first
- Contacting journalists
- Public rebuttals
- Legal allegations
Hard limits
- No contact without approval.
It stops when
- Done: case closed or drafted.
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