Google Is Now Top of Funnel. AI Is Where the Real Buyers Are.

There's a narrative doing the rounds that AI is killing web traffic. The panicked headlines about collapsing Google referrals, the dire Gartner predictions, the scary charts showing publisher traffic in freefall. And the numbers are real. Google Search referral traffic fell 34% in a single year. Small publishers have lost up to 60% of their search traffic over two years.

But this narrative misses the most important shift happening in digital marketing right now. It's not about how many visitors you're getting. It's about where those visitors sit in the buying journey when they arrive.

The data is unambiguous: visitors arriving from AI chatbots are dramatically more qualified, more engaged, and more likely to convert than visitors from traditional search. Google has quietly become a top-of-funnel awareness channel. AI is emerging as the most valuable bottom-of-funnel source most businesses have never optimised for.

The Funnel Has Flipped

Think about what happens when someone uses Google. They type in a query. They scan ten blue links. They open three or four tabs. They skim, compare, bounce, and repeat. They're browsing. They're in discovery mode. They might be ready to buy, or they might just be killing time. You don't know, and neither does Google.

Now think about what happens when someone asks ChatGPT or Claude a question like "What's the best WordPress booking plugin for a small yoga studio?" or "Which email marketing platform integrates best with WooCommerce for a shop doing under £5k a month?"

The AI doesn't return ten options and wish them luck. It synthesises information from across the web, evaluates it against the specific criteria in the question, and delivers a curated recommendation. If the user then clicks through to your website, they've already been filtered, compared, and pre-qualified by the AI. They know what they want. They know why they're on your site. They're not browsing. They're buying.

That's the fundamental difference, and the conversion data proves it.

The Conversion Gap Is Enormous

A 13-month study analysing LLM referral traffic across multiple brand sites found an 18% conversion rate from AI-referred visitors. To put that in context, typical organic search conversion rates sit between 2% and 5%.

Semrush's research quantifies it differently but reaches the same conclusion: the average LLM visitor is worth 4.4 times more than the average traditional organic search visitor based on conversion rates.

Knotch's analysis across enterprise clients found that LLM traffic accounts for just 0.13% of total visits but drives 0.28% of conversions, punching at more than twice its weight. Half of the brands in their study saw LLM conversion rates that were more than double their sitewide averages. Some saw triple-digit gains.

And it's not just about the final conversion event. The engagement data tells the same story. LLM visitors spend 1.9 times longer on page and scroll 1.3 times deeper than the sitewide average. These aren't casual browsers. They're people who arrived with purpose and are consuming your content with intent.

Meanwhile, LLM conversions grew 124% in a recent four-month window. Over the same period, organic search conversions fell 38%.

Why AI Traffic Is Inherently Bottom-of-Funnel

The reason LLM traffic converts so much better isn't a mystery once you understand how AI discovery works compared to traditional search.

The AI does the comparison shopping for them. When a user asks an AI to recommend a solution, the model searches across multiple sources, evaluates options against the user's specific criteria, and presents a synthesised answer. The user doesn't need to open twelve tabs and read twelve "Top 10" listicles. By the time they click through to your site, the evaluation phase is already complete.

The questions are more specific. Google search queries tend to be short and broad: "best WordPress SEO plugin", "email marketing tool", "online booking system". LLM prompts are longer and more contextual: "What's the best way to set up automated abandoned cart emails on a WooCommerce store that also syncs with my Mailchimp audience segments?" Longer, more specific questions signal higher intent. Users who frame questions this way know what they need and are looking for a specific solution, not a general education.

AI pre-qualifies based on fit, not ranking. Google ranks pages based on authority signals, backlinks, and keyword relevance. An AI recommends solutions based on fit to the user's stated needs. A page ranked number one on Google might be irrelevant to a specific use case. An AI recommendation, by contrast, has already assessed whether your offering matches what the user is asking for. This means the traffic that arrives is pre-filtered for relevance.

The trust transfer is different. When a user clicks a Google result, they're choosing from a list and evaluating your credibility themselves. When an AI recommends your product or service, there's an implicit endorsement. The user has delegated their research to a system they trust, and that system pointed them to you. They arrive with a baseline of confidence that a Google click doesn't provide.

Follow-up prompts keep users inside the AI. Research shows that users' opening questions trigger web searches within AI platforms, but follow-up prompts rarely do. This means the click-through that does happen is deliberate. The user has received the AI's synthesised answer, decided they want to go deeper, and specifically chosen to visit your site. That's the opposite of a casual SERP click.

What This Means for Your Marketing Strategy

If you accept the premise that Google is top-of-funnel and AI is bottom-of-funnel, the strategic implications are significant.

Google Ads become your awareness play. Paid search still works for getting into the consideration set. Brand awareness, category education, and top-of-funnel visibility remain Google's strength. But the expectation of direct conversion from a Google click needs to be recalibrated. You're planting seeds, not closing deals.

AI visibility becomes your conversion play. The visitors who arrive from LLM recommendations are your most valuable prospects. Optimising your site for AI citation and recommendation is not an SEO side project. It's a direct investment in your highest-converting acquisition channel.

Content strategy needs to split along funnel lines. Your Google-facing content should focus on awareness and education: what is this category, why does it matter, what are the key considerations. Your AI-facing content should focus on specificity and authority: detailed product information, comparison data, unique datasets, and clear answers to specific use-case questions. These are different content types serving different stages of the journey.

Measurement needs to evolve. If you're still measuring success purely by traffic volume, you're optimising for the wrong metric. A 50% drop in Google traffic coupled with a 200% increase in AI referral conversions is a net win for your business, but your analytics dashboard won't tell you that unless you're tracking AI referral sources separately.

The Practical Reality

Let's be clear about the current scale. AI referral traffic is still small in absolute terms. It accounts for around 1-2% of total referral traffic for most sites. ChatGPT drives the overwhelming majority of it at 87.4% of AI referrals, with Perplexity, Gemini, and Copilot splitting the rest.

But the trajectory is undeniable. AI referral traffic grew 527% year-on-year in early 2025. Semrush projects LLM traffic will overtake traditional Google search by the end of 2027. Three in four Americans already search with AI weekly.

The businesses that build AI visibility now are the ones that will own this channel when it reaches maturity. The ones that treat it as a rounding error will find themselves competing for an ever-shrinking share of traditional search clicks, against an ever-growing number of competitors, in a channel that delivers progressively less qualified traffic.

What Makes a Website Convert AI Traffic

Understanding that AI delivers better leads is only useful if your site is set up to receive them. The visitors arriving from LLM recommendations have different expectations to those arriving from Google.

They expect the answer immediately. They've already had a summarised overview from the AI. They're clicking through for depth, confirmation, or a specific action (sign up, get a quote, start a trial). If your landing page opens with generic waffle before getting to the substance, they'll leave. Lead with value.

They're comparing against what the AI told them. The AI has already described your offering in its recommendation. If your site contradicts that description or doesn't match the positioning the AI gave, you'll lose trust. Ensure your site content is accurate, current, and consistent with how you want AI systems to describe you.

They want to act, not research. These visitors are past the research phase. Clear calls to action, straightforward pricing (where applicable), and frictionless conversion paths matter more for AI traffic than for any other channel. Remove barriers.

The Bottom Line

The shift from search to AI discovery isn't a story about losing traffic. It's a story about the marketing funnel being restructured by technology.

Google is becoming the channel where people discover that solutions exist. AI is becoming the channel where people decide which solution to choose. The traffic from AI is smaller in volume, but it carries dramatically more commercial intent and converts at multiples of what traditional search delivers.

The businesses that recognise this shift and optimise for it will capture the most valuable traffic source of the next decade. The ones that keep chasing Google traffic volume as their primary metric will be optimising for a channel that's increasingly top-of-funnel, while their competitors quietly own the bottom.

The question isn't whether this shift is happening. It's whether you'll be visible when your next customer asks AI for a recommendation.

Sources: Chartbeat/Axios (March 2026), Graphite (March 2026), Search Engine Land (February 2026), Knotch, Semrush, Ahrefs, Conductor AEO/GEO Benchmarks Report (January 2026), SE Ranking, SparkToro, Previsible AI Traffic Report (2025), Gartner, Bain & Company, Nectiv, Profound.