Which mentions actually line up with showing up in AI answers?
Get the 5 AI search links every Tuesday
Hi from Tokyo! On Dec 9 I'm speaking at AIBB Tokyo with speakers from Harvard and SoftBank. Come say hi if you're around.
Most teams chase mentions in big publications to show up in AI answers. Ahrefs looked at 10,000 brands this week and found a mention on a small niche site counts about as much, and Reddit threads count the most. Yet 29% of brands had none.
Cheers,
Tanya
P.S. I respond to every email.
This week in 30 seconds
- Which mentions actually line up with showing up in AI answers? (Ahrefs)
- How big is ChatGPT search compared with Google, really? (Search Engine Journal)
- Is ChatGPT about to get its own Search Console? (Search Engine Roundtable)
- Why might your rank tracker be showing Google links instead of your pages? (Search Engine Journal)
- Is there a free checklist for Google's content quality rules? (Aleyda Solis)
Which mentions actually line up with showing up in AI answers?
Ahrefs · Nov 3
Source: ahrefs.com, Which Types of Mentions Correlate the Strongest With AI Visibility? (Louise Linehan), Nov 3 2026.
- Ahrefs compared 7 kinds of mentions for 10,000 US brands across about 260 million AI prompts.
- Reddit mentions in ranking threads lined up best with AI visibility (0.75).
- A mention on a small niche site counted about as much as one on a big publication.
- 29% of brands had no Reddit threads mentioning them at all.
Why it matters: It is correlation, not proof, but it shows where the cheapest wins are.
What to do: Find the Reddit threads that rank for your category and check if your brand is in them.
How big is ChatGPT search compared with Google, really?
Search Engine Journal · Nov 3
Source: searchenginejournal.com, Google's Search Volume Is About 5x ChatGPT's (Matt G. Southern), Nov 3 2026.
- OpenAI's ChatGPT product lead said ChatGPT gets about one search-like query for every five on Google.
- He did not say which Google number, period or regions he used.
- In Ahrefs' traffic tracker for October, ChatGPT sent 0.36% of referral traffic. Google sent 24.74%.
Why it matters: Share of questions and share of visits are not the same thing.
What to do: Track both: AI answers shape your shortlist, Google still sends most of the visits.
Is ChatGPT about to get its own Search Console?
Search Engine Roundtable · Nov 6
- OpenAI is reportedly building a way to claim your business in ChatGPT, according to a report relayed by Search Engine Roundtable.
- It would show how often you are mentioned and clicked, day by day.
- Reportedly no prompt data and no competitor benchmarks. No launch date yet.
Why it matters: It would be the first ChatGPT visibility data that comes from OpenAI itself.
What to do: Know which login proves you own your brand (domain, store or business email) so you can claim it fast.
Why might your rank tracker be showing Google links instead of your pages?
Search Engine Journal · Nov 5
Source: searchenginejournal.com, Google Appears To Limit Goto Result Links (Matt G. Southern), Nov 5 2026.
- SEJ reports that Google appears to be slowing down how fast its google.com/goto result links can be followed.
- DataForSEO, Serper and SerpApi all reported problems from about Oct 30.
- Some rank data saved a Google redirect instead of your real page.
Why it matters: A page that seems to vanish from your report may still be ranking.
What to do: Before you panic about a lost URL from late October, check if the stored URL is a google.com/goto link.
Is there a free checklist for Google's content quality rules?
LinkedIn (Aleyda Solis) · Nov 3
- Aleyda Solis updated her E-E-A-T, helpfulness and page experience checklist with Google's latest criteria.
- It is a free Google Sheet, linked in her comment.
- She also added a Claude skill that helps you run the assessment.
Why it matters: The pages that hold up in Google are usually the ones AI tools pick up too.
What to do: Run your top 3 product pages through the checklist and fix the first gap it finds.
That's this week. Which one should I go deeper on? Just reply.