Before a B2B buyer reaches a vendor’s website today, they may have already compared features, evaluated pricing, and built a shortlist — all inside an AI platform. By the time they follow an AI citation to a product page, the selection work is largely done.
That’s the insight behind a growing body of data showing AI-referred website visitors convert at much higher rates than visitors from other channels. It’s not that AI sends better traffic. It’s that AI sends buyers who are already ready.
Understanding why starts with looking at how people actually use AI.
According to a study by OpenAI, Duke University, and Harvard University, nearly half of all ChatGPT interactions are “Asking” prompts, where the user is seeking information or clarification to inform a decision — requesting vendor comparisons, recommendations, and evaluations. That’s consideration-stage B2B buyer behavior. When a buyer uses AI to evaluate vendors and follows a citation to a product or pricing page, they arrive having already done the shortlisting. They’re conversion ready.
The same study also classifies approximately 40 percent of prompt behavior as “Doing,” where the user wants to produce some output or perform a particular task — predominantly writing. Another 11 percent is classified as “Expressing,” where the user shares views or feelings without seeking information or action.
This leads to the zero-click effect: Doing and Expressing prompts are being answered within AI platforms themselves, which is partly why website traffic is declining. AI is satisfying user intent in-platform, without sending traffic back to the source.
AI Referral Traffic: What the Data Shows
AI referral traffic refers to website visits that originate from AI platforms such as ChatGPT, Perplexity, or Google Gemini. When these platforms cite a source in response to a user query, users who click through are counted as AI-referred visitors — and the data on those visitors is striking.
Several studies point in the same direction:
- AI-referred visitors convert at 4.4 times the rate of traditional search visitors.
- LLM-referred traffic converts at 18 percent — the highest rate of any measured source.
- ChatGPT-sourced B2B leads close at just over 4 percent, nearly double the next best channel.
Top-of-Funnel Traffic: What the Data Shows
Top-of-funnel traffic refers to website visits driven by awareness-stage content, such as blog posts, explainer articles, and educational resources that attract buyers who are early in their research, before they have identified a specific vendor or solution. Historically, this traffic has been the primary measure of content marketing reach and organic search performance for B2B brands.
Simultaneously, that traffic is under pressure:
- HubSpot’s organic traffic dropped sharply from 13.5 million in November to 8.6 million in December, with its top-of-funnel B2B blog bearing the brunt of the decline.
- Semrush found that 88 percent of queries triggering AI Overviews are informational in intent — the type that has historically driven awareness-stage web traffic.
- PAN’s report on the state of brand visibility in a zero-click world revealed that 38 percent of people of people skipped visiting a website when an AI summary answered their question.
These trends are connected. Understanding why requires a deeper dive into AI user prompt behavior.
Related Read > > PAN’s AI Overview Tracker
Zero Click: Why Top-of-Funnel B2B Content Is Losing Traffic
A zero-click result occurs when a search engine or AI platform answers a query directly, eliminating the need for the user to visit any website. This is increasingly common with AI Overviews and conversational AI platforms — and it’s a key mechanism behind declining top-of-funnel traffic.
When a “Doing” or “Expressing” user prompts AI to draft an email, build a project plan, or write a script, they get the output they need without leaving the conversation. No click. No website visit. Top-of-funnel content is still informing AI answers — but the value is being captured by the AI platforms themselves, which serve that content to users without passing traffic back to its source.
AI Transfers Ready Buyers: Why AI Referral Traffic Converts at Higher Rates
What suppresses awareness-stage traffic is precisely what makes consideration-stage AI traffic so valuable. An “Asking” user querying AI to compare product features or evaluate vendor pricing cannot get everything they need from a conversational response based on training data alone. They need current information to guide their decision, so the AI cites sources — and those sources are often product pages, pricing pages, and comparison content.
The buyer who follows that citation has already done the selection work. They know the category. They’ve applied their constraints. They’ve compared alternatives. By the time they land on a vendor’s page, they’re ready to act. It isn’t that AI traffic is inherently more valuable. It’s that the AI user prompt behavior producing that traffic precedes a purchasing decision almost by definition.
B2B Brand Perception: How AI Is Changing the Way Buyers Shortlist
There’s a consequence of this dynamic that goes beyond traffic and conversion rates. When AI responds to a consideration-stage “Asking” prompt, it doesn’t just (or always) cite sources. It summarizes, compares, and frames vendors in-platform before the potential buyer clicks anywhere. That summary is often the first substantive perception a prospect has of a B2B brand.
The accuracy of that summary depends largely upon the consistency and clarity of the information the AI has absorbed about a brand across the web. When the language a vendor uses on its website differs from how it’s described in its social presence, press coverage, or third-party reviews, the AI synthesizing that information may fill gaps with hallucinations — generating plausible-sounding but inaccurate descriptions of a brand’s products, positioning, or differentiators.
This is something the team at PAN encounters regularly when auditing a B2B brand’s AI visibility. Small inconsistencies surface that undermine credibility, making messaging alignment across every channel a prerequisite for AI platforms to interpret a brand as one unified, authoritative source.
This matters at the consideration stage specifically because buyers using Asking prompts are actively forming a vendor shortlist. A vendor whose AI-generated summary is vague, outdated, or inconsistent with its actual positioning is likely being passed over at exactly the moment a buyer is deciding who makes the list. The sales team never gets the chance to correct that impression, because the buyer has already moved on.
Key Takeaways: What B2B Brands Can Do About It
Accept that top-of-funnel web content is losing traffic but has not lost its purpose. AI draws on it to shape perception at a layer that web analytics don’t capture. The marketing value remains awareness — and most buyer journeys are non-linear. The goal is to have your content present wherever your target audience is active and receptive, while ensuring that content is genuinely useful and substantive.
Realize that product, pricing, and comparison content is now doing two jobs. It converts human visitors and earns AI citations that bring more of them. Websites with poor information architecture and pages with indistinct feature descriptions, gated pricing, and undifferentiated messaging are failing at both jobs. Specificity and clarity are no longer just good UX — they directly affect whether AI includes a vendor in a recommendation.
Embrace integrated, consistent messaging as a competitive requirement. From B2B web design to content marketing and public relations, PAN’s services support the approach AI platforms need to interpret your brand as one unified, authoritative source. Contact us to learn more.
