
There are three new studies out over the past week that suggest that spending time, money and resources on AI visibility and AI search may not be worth it. In short, the studies all show that AI Search or AI services send a tiny fraction of traffic (and conversions) when compared to Google Search.
In short, AI Search sends almost no traffic and conversions when compared to other traffic drivers on the web.
Here are the studies:
- GEO is a distraction from Marie Haynes
- How Much Traffic Do AI Assistants Actually Send, And Does It Convert Better? from SEO Works
- Does AI drive more direct traffic? from Kevin Indig
Marie wrote, "I looked at 12 months of GA4 data across 100 different websites of varying sizes with over 145 million sessions. In this dataset, sessions from ChatGPT amounted to just 0.62% of the volume driven by Google organic."
She added, "However, even though ChatGPT visitors converted at more than 4.5x the rate of Google organic visitors, Google still drove 96.5% of the total conversions across all channels analyzed." Also while ChatGPT visitors converted better at 13.8% vs 3.02% for Google organic, Google still drove 27x more conversions than all the other LLMs combined.
She showed this visually in this graphic:
SEO Works found that organic search delivered 39 times more traffic than AI assistants and for every 100 conversions organic search delivered, AI referrals delivered 3.2. It also said, "AI referral traffic converts worse against paid search (0.67x) and better against direct traffic (2.48x)."
Here is a chart from SEO Works:
Then Kevin Indig wrote, "I found that the amount of Direct traffic to longtail pages in my sample driven by AI is not detectable yet. We also can’t yet find a reliable relationship between AI Visibility and branded search or direct traffic. Some companies with outstanding AI growth might be able to detect this, but across 12 sites of different sectors and sizes, I couldn’t see it."
Rand Fishkin did a great job summing all this up in a video on LinkedIn - here is that video:



