The problem
Your website says you can 'cure' or 'fix' something, and a complaint to the advertising regulator or your own regulator follows. If you publish 8 pages and posts a month and checking each against the rules takes 10 minutes, that's 80 minutes a month.
What the AI does
Paste your own draft web page or post into ChatGPT, Claude or Gemini and ask it to flag claims about conditions, outcomes or cures that would need evidence under advertising rules. You then check each flagged line against the ASA and CAP code and your regulator's guidance (GOsC or the GCC) and rewrite it. The assistant is a second pair of eyes and cannot certify anything.
A worked example
A chiropractor in Glasgow publishes about 8 pages and social posts a month. He pastes each draft into an AI assistant and asks it to flag claims that need evidence.
Before
10 minutes of manual checking per item, 80 minutes a month, easy to skip
After
About 5 minutes to review each flag, 40 minutes a month, with a consistent first pass
What it's worth
About 40 minutes a month saved (5 minutes x 8 items), worth around £33 if your time is worth £50 an hour. The larger value is avoiding a complaint, for which there's no reliable figure.
How to set it up
- Write your page or post as you normally would.
- Paste it into your AI assistant and ask it to list any claims about conditions, cures or outcomes that would need evidence.
- Check each flagged line against the ASA and CAP code and your regulator's advertising guidance.
- Rewrite or remove anything you can't support with evidence.
- Keep a note of the evidence behind each claim you keep.
What goes wrong
- The assistant can miss a claim or tell you something is fine when it isn't. You remain responsible for what you publish.
- It may quote US or out-of-date rules. Use the ASA, CAP, GOsC and GCC guidance themselves.
- Never paste patient testimonials with names or case details.
- Passing the check doesn't make a claim true.