This one comes up in nearly every CMO conversation I have: "We're not showing up in ChatGPT. What do we fix?"
The hard part is diagnosis. Teams default to "we need more content" when the real problem is somewhere else. Here are the five I see, with proof from practitioners publishing real results and a quick test for each.
Technical: AI can't read the page
Vercel and MERJ found that none of the major AI crawlers (ChatGPT, Claude, Perplexity) run JavaScript. If your copy loads through JavaScript, those bots see an empty page.
Copy: the answer is buried
Kevin Indig's research shows ChatGPT pulls most citations from the top 10 to 20% of a page. Seer Interactive restructured 11 existing articles to answer the question up front and grew AI Overview citations 344% in five months.
Positioning: you aren't the obvious answer
Josh Blyskal at Profound found AI passes over "schedule a demo" pages in favor of pages that actually answer the question. Fire&Spark built one piece around a client's own clinical trial data: 275 AI citations in 30 days, first on 14 of 20 tracked prompts.
Off-page: others aren't talking about you
Across 75,000 brands, Ahrefs found brand mentions on other sites track AI Overview visibility far more closely than backlinks do. When Foundation asked ChatGPT about 15 SEO Week sponsors, 58% of the sources it cited were content those brands didn't own.
Consistency: your facts don't match
When sources disagree or stay vague, AI goes with whatever is most specific, right or wrong. MIT Sloan Management Review described a company that ChatGPT said had five funds instead of seven, and it got their fees wrong too.
Why this is a team problem
Each of the five has a different owner: dev, content, product marketing, PR, ops. That's why "AI visibility" often belongs to no one, and why the upskilling question matters more than the tool question.
Sources: Vercel and MERJ crawler study; Kevin Indig, Growth Memo; Seer Interactive; Profound; Fire&Spark; Ahrefs brand-mention study; Foundation; MIT Sloan Management Review.
