Demandbase reports that visits arriving at B2B websites from ChatGPT grew from about 645,000 a month in June 2025 to 2.6 million in June 2026. That’s a 303% increase, and it’s the clearest figure I’ve seen on how quickly B2B buyers are being sent to websites by AI tools. It’s also a vendor’s own data from its own customers, so it needs reading with some care.

See where your pipeline is actually breaking. The B2B diagnostic scores your positioning, demand generation approach, and lead quality signals against what high-performing teams do differently.

Run the B2B diagnostic

What Demandbase found

Demandbase published the figures on 12 August 2026, using data from its customer platform between June 2025 and July 2026. The headline points:

  • ChatGPT-referred visits to the B2B websites it measures nearly quadrupled over the year, from about 645,000 a month to 2.6 million.
  • ChatGPT accounts for the large majority of AI-referred traffic in the data. Perplexity referrals declined, and Gemini and Claude stayed flat. The release gives no figures for those three, so I wouldn’t read much into them.
  • The sample covers more than 11 billion website visits across 1,584 Demandbase customer accounts.

"Referral traffic" here means someone clicking a link in an AI assistant and landing on a website.

What it doesn’t tell you

I’d keep these in mind before building a strategy on it.

  • It’s vendor data. Demandbase is describing its own customers, and the release isn’t independent research.
  • The sample isn’t the whole market. These are B2B companies that pay for Demandbase, which tends to mean larger businesses than many of the firms I talk to.
  • It counts visits, not leads. The release doesn’t report conversion rates, lead quality or pipeline. More visits from ChatGPT doesn’t show those visits turn into enquiries.
  • It only counts clicks it can trace. Someone who asks ChatGPT for suppliers, then searches for your name or types your address in later, won’t appear as a ChatGPT referral. So the real influence is probably larger than the visits show. That’s my inference, not something the release says.

What this means for a smaller B2B business

The numbers don’t prove AI tools are about to replace search. They do suggest that being the page an AI tool points to is now worth some of your attention. For most B2B businesses, that comes down to a few plain things:

  1. Answer real questions clearly. AI tools tend to point to pages that answer a specific question plainly and early. A page that opens with the answer is easier to quote than one that builds up to it. I covered the basics in what AI search is.
  2. Say who you are and who you work with. If the page doesn’t make clear what you do, where you’re based and who for, there’s little for a tool to cite.
  3. Keep the facts consistent. If your site says one thing and your LinkedIn page and directory listings say another, that makes you harder to describe accurately.
  4. Be findable where people check. I wrote about LinkedIn and AI search citations for exactly this reason.

What I’d check on your own site this week

This takes about ten minutes and tells you more than any industry figure.

  • Open your analytics and look at where your traffic comes from. Check whether any visits are listed as coming from chatgpt.com or similar AI sources. Many sites will find a small number, and some will find none.
  • Note which pages those visits land on. That shows what’s already being cited.
  • Ask ChatGPT, Perplexity and Google a few of the questions your buyers ask. See whether your business or your pages appear, and what they say about you.
  • If they get something wrong, fix the page or profile that’s the likely source.

That’s useful, because you can fix what you can see.

The short version

AI referrals to B2B sites are growing quickly according to one large vendor’s data, with the caveats above. The sensible response is a modest one: make your pages easy to understand and quote, keep your details consistent, and measure what’s actually arriving before changing anything big.

If you’d like a structured view of how visible and clear your marketing is right now, the free diagnostic scores it across the areas that matter, and I’m happy to talk through the results on a call.

Common questions

Is ChatGPT traffic going to replace Google for B2B?

This data doesn’t say that. It shows ChatGPT referrals growing from a small base, within one vendor’s customer set. Search and AI tools sit side by side for most buyers at the moment.

Does AI referral traffic convert better?

The Demandbase release doesn’t report conversions, so I wouldn’t claim it does. Check your own analytics before assuming either way.

How can I tell whether ChatGPT is sending me visitors?

Look at the referral or source report in your analytics for chatgpt.com or similar. If nothing appears, that’s a useful baseline.

Source: Demandbase, press release on ChatGPT referrals to B2B websites, 12 August 2026.

See where your pipeline is actually breaking. The B2B diagnostic scores your positioning, demand generation approach, and lead quality signals against what high-performing teams do differently.

Run the B2B diagnostic