AI News/AI News

Under 7% of ChatGPT chats show purchase intent, SparkToro panel finds

SparkToro's analysis of a U.S. opt-in panel offers a benchmark for marketers weighing ChatGPT ads and AI visibility budgets.

Key points

  • SparkToro analyzed 41,961 ChatGPT conversations from 12,321 U.S. panelists, supplied by Verso, and found 6.9% showed purchase intent.
  • 87.9% of ChatGPT responses in the weighted sample named no brand, product or business; 12.1% mentioned at least one.
  • Panel skews toward mobile, younger users, women and lower incomes, and excludes API, most desktop and work use, so it is not a usage count.
  • Shopping, brands and local services made up 5.3% of conversations; image and visual design requests led at 17.6%.

Why it matters for marketers: Teams weighing ChatGPT ads or AI visibility spend get a measured benchmark for buying intent, but should check it against their own referral and ad data first.

SparkToro analyzed ChatGPT conversation histories from a U.S. opt-in panel run by Verso and found purchase intent in 6.9% of conversations. The sample covered 41,961 conversations from 12,321 people between Jan. 1 and Oct. 29, 2026, with most figures drawn from a weighted sub-sample matched to U.S. demographics.

Within that group, 4.9% of conversations involved someone planning to buy and still comparing options, and 2.1% involved questions about a specific brand, product or service. Most prompts (86.7%) did not ask about brands at all, and 87.9% of responses named none. Rand Fishkin, SparkToro's co-founder and CEO, wrote, "I suspect this is bad news for ChatGPT's advertising opportunities."

For marketers, the number frames how much of ChatGPT use is commercial. Fishkin concluded that AI visibility matters mainly because of Google's AI Overviews and AI Mode, and he questioned whether tracking ChatGPT visibility pays off unless Bing rankings are tracked too. He also pointed to earlier SparkToro Google search data, though the categories differ and the percentages cannot be compared directly.

The limits are real. The panel skews younger, female and lower income, was recorded mostly on mobile, and excludes API requests, so desktop and work use may be underreported. Claude coded the conversations, and a second coder agreed on categories 90% of the time. No separate agreement figure was reported for purchase intent. SparkToro sells audience research software.

For PPC teams looking at ChatGPT Ads, the figure is a snapshot of conversation mix, not a count of buying conversations across ChatGPT. The practical step is to compare it with your own ChatGPT referral and ad data before it goes into a budget plan.

Sources

This brief is our own summary. Read the full reporting at the source.

You read about AI automation. We run it.

DocFluence, built by Z Media, writes, designs, films and publishes for your brand every day, in your voice, from one login.