Drafts that show where every claim came from.
Drafts posts and newsletters from your approved material, with a source on every claim. Anything it can’t support gets flagged.
Where a person steps in: someone approves the draft, handles anything flagged and publishes. The agent can’t publish.
Details: drafts that show where every claim came from
Why it matters
Someone drafts from scratch. Someone else checks the facts if there’s time. The approval chain catches what it can. Errors can surface after publishing, when they’re hardest to fix.
How it runs
- A request comes in for a post, a newsletter or an announcement.
- The agent drafts from your style guide, your program and service facts, and past approved pieces.
- It marks each factual claim and its source, and marks any figure it can’t source as unknown.
- The draft enters your approval chain at the start, not after the fact. A person approves and publishes.
Controls
Can
- Read your approved source material and style rules.
- Draft and flag.
Can’t
- Publish, post or schedule anything on any channel.
Is instructed to
- Mark any figure it can’t source as unknown, and a person makes the decision before approval.
- Flag any claim it can’t support rather than soften it.
What it needs from you
A named owner. Your style guide or brand rules. A written set of facts the agent may draw on. A map of who approves what before it goes public.
How we’d measure it
Claims flagged before approval compared with claims corrected after publishing, and the time from request to an approved draft. We agree the baseline before we build.
First step
The Agent Readiness Assessment, with its marketing and communications module.