Does AI Visibility Actually Win Customers?

Rand Fishkin's article on AI visibility questions the relationship between the effort brands put into appearing in AI answers and the results they get. The research he cites shows that AI tools still send relatively little traffic to websites. Traffic volume alone, however, does not tell us what those visits are worth to a business.

Getting a brand recommended by ChatGPT, bringing someone to its website, receiving a form submission and receiving a payment are separate events. Unless we know how these events relate, greater visibility alone is not evidence of customer acquisition.

Less traffic, higher conversion rates?

In The SEO Works' study of 95 websites, organic search brought far more traffic. Across the 63 sites included in the conversion comparison, AI referrals converted at a higher rate overall. At 23 of those sites, the result went the other way.

The study measured GA4 conversion events from September 2025 to August 2026; those events do not necessarily represent sales. It could not isolate the effect of Google AI Overviews, and the organisation publishing the research sells SEO services. Interpreting the result requires knowing which conversion was measured.

A different result at the deposit stage

In three months of data from a healthcare company I work with, fewer than 1% of leads came from ChatGPT. We track whether these leads reach the quote stage and make a deposit payment.

The table compares each source with ChatGPT. Each rate uses that source's own leads as its denominator.

On the lead-to-quote rate alone, ChatGPT appears to outperform Google Ads. By the lead-to-deposit rate, Google Ads comes out ahead.

Attribution does not show whether an investment paid off

A lead recorded as coming from ChatGPT does not establish that a particular investment in AI visibility generated it. Tracking the source attributed to a lead and measuring the additional results an investment produces are different tasks.

When I previously examined what an AI visibility score measures, I focused on the decisions that measurement could support. This time, we have actual leads rather than a visibility score. We still need to be clear about which questions the data can answer.

ChatGPT's small share of total leads does not make it pointless to track. Similar lead-to-quote rates do not tell us how lead-to-deposit rates will compare, either. And neither rate accounts for the cost of each source. These figures cannot tell us which channel is more profitable.

To understand whether AI visibility brings in customers, we need to look beyond the visibility report. In our example, lead-to-quote rates alone would have made ChatGPT and organic traffic look similar. Including lead-to-deposit rates revealed a difference. We cannot judge a channel's performance without specifying the outcome.