Flat AI search traffic with rising enquiries means AI is sending fewer, better-qualified visitors. Judge it on leads, not sessions.
If your AI search traffic looks flat but enquiries are going up, nothing is broken. AI assistants answer the research questions themselves and send fewer, better-qualified visitors, so sessions are now a poor measure of whether AI search is working. Leads are the better one.
We see this pattern across a lot of the brands we work with. Organic sessions hold steady or dip slightly, the traffic graph in the monthly report looks like a shrug, and yet the sales team is busier than it was a year ago. When you ask new enquiries how they found the business, more and more of them say ChatGPT, Gemini or Perplexity.
This matters because the traffic graph is still the first thing most boards look at. Judge AI search on sessions alone and you’ll underrate the channel that is quietly doing the most selling.
Why AI search sends less traffic but better leads
AI search sends fewer visitors because the assistant does the early research for the buyer. Someone who once read five comparison articles now asks one question and gets a shortlist. By the time they click through to your site, they’ve already compared options, read a summary of what you do and decided you’re worth talking to.
Google’s own results work the same way now. Pew Research Center tracked real browsing behaviour and found people clicked a traditional result on 8% of visits when an AI summary appeared, against 15% when there wasn’t one (Pew Research Center, July 2025). Fewer clicks happen, but the ones that do come from people further along.
Ahrefs saw the other side of this in its own data. AI search brought in 0.5% of its visitors but 12.1% of its signups over the period it studied (Ahrefs). One company’s numbers won’t match yours exactly, but the direction is the same one we see on B2B accounts.
What a flat traffic graph is hiding
A flat traffic graph usually hides a change in who is visiting, not a lack of interest. Three things tend to be going on underneath it:
- Fewer learners, more buyers. Top-of-funnel “what is” searches are increasingly answered inside the AI result. The people who still click are closer to a decision.
- Brand searches after an AI mention. A buyer sees your name in a ChatGPT answer, then searches for you directly or types your URL. Analytics records that as direct or branded search, not as AI.
- Visits your analytics can’t label. Some AI apps don’t pass a referrer, so those sessions land in “direct” with nothing to show where they came from.
Put together, the AI channel is often bigger than the referral report suggests, and its value shows up in the pipeline rather than the sessions chart. We covered the wider version of this problem in why more traffic isn’t fixing your lead problem.
How to measure AI search properly
Measure AI search the way you’d measure a salesperson: on the conversations it starts. This is the set-up we put in place for clients:
- Ask every enquiry how they found you. Add a free-text “How did you hear about us?” field to your forms. Free text beats a dropdown, because people write “ChatGPT recommended you” in their own words.
- Group AI referrers in analytics. Create a custom channel group in GA4 for chatgpt.com, perplexity.ai, gemini.google.com, copilot.microsoft.com and similar sources, so AI visits stop hiding inside “referral”.
- Track enquiries and pipeline by channel, not sessions. Report form fills, booked calls and won deals against each source. A channel with a tenth of the traffic and a third of the leads should get attention, not cuts.
- Watch branded search and direct traffic. A steady rise in people searching your company name is often the clearest sign AI tools are mentioning you.
- Check how often AI names you. Run the questions your buyers ask through the main assistants each month and record whether you appear. This is the idea behind share of model, and it tells you whether you’re in the answer before the leads ever arrive.
If you want to see where you stand today, our free AI visibility check runs the basics for you.
Which numbers to put in front of your board
Put leads, pipeline and cost per lead at the top of the report, with traffic underneath as context. A useful monthly view looks like this:
| Metric | What it tells you |
|---|---|
| Enquiries from AI and organic search | Whether search is bringing in business |
| Self-reported source on forms | How many buyers found you through an AI assistant |
| Branded search volume | Whether more people know your name |
| Share of AI answers for your key questions | Whether you’re recommended before anyone clicks |
| Organic sessions | Context only: a dip with rising leads is a good sign |
This is the same thinking behind what a good SEO agency report should show you every month. If your current report leads with traffic and stops there, it is answering the wrong question.
What this means for your SEO budget
It means you should judge AI search on what it earns, not on how busy the site looks. High-value B2B deals make this easy to work out: if one extra client covers months of work, the channel only needs to bring in a handful of the right enquiries to pay for itself. That calculation is far more useful than a sessions target, and it’s how we suggest B2B companies think about SEO timelines too.
It also changes what content is worth making. Some pages earn their keep by being the source an AI tool quotes, even if they never rank in the top three for a big keyword. Clear answers, proof such as case studies and reviews, and consistent information about your business across the web all help you get named in those answers.
If you’d like help setting this up, our AI SEO agency team and our B2B SEO specialists can show you which channels are bringing in your leads. Book a free strategy session and we’ll walk you through it.
Related reading: our guides to tracking AI Overviews and why AI cites review sites instead of your own website.
Common questions
Why is my traffic flat when my leads are going up?
AI assistants answer early research questions themselves, so fewer people click through. The ones who do are further along and more likely to enquire, so leads can rise while sessions stay flat.
How do I see AI search traffic in GA4?
Create a custom channel group that collects referrers such as chatgpt.com, perplexity.ai, gemini.google.com and copilot.microsoft.com. Some AI apps don’t pass a referrer, so also ask enquiries how they found you.
Does AI search traffic convert better than Google traffic?
Often, yes. Ahrefs found AI search brought 0.5% of its visitors but 12.1% of its signups. Results vary by business, so track enquiries by source on your own site.
What should I report instead of traffic?
Report enquiries, pipeline and cost per lead by channel first, then branded search and how often AI tools name you. Keep sessions as context underneath.
