AI Visibility · 25 August 2026

By Eddie Eastwood, Founder

Does AI Visibility Actually Drive Real Business Leads?

Does AI visibility actually drive real business leads?

Yes, for some businesses, right now, it does. ChatGPT and other AI platforms send real visitors to real business websites, and a share of those visitors convert, the same way a share of Google's organic visitors always has. What's genuinely unresolved is the size of that effect for any specific business and how cleanly it can be measured. This page sets out what's measurable, what isn't yet, and how to weigh the investment.

What the traffic data actually shows

AI platforms are still a small slice of referral traffic, but the slice is growing fast. Similarweb tracked the top 1,000 websites worldwide and found AI platforms sent over 1.13 billion referral visits in June 2025, up 357% from June 2024, against Google Search's 191 billion referrals in the same month; ChatGPT accounted for more than 80% of that AI traffic. Ahrefs separately studied 3,000 websites and found 63% had received at least one AI-referred visit, with ChatGPT sending just over half, Perplexity close to a third, and Gemini around 18%. Ahrefs flagged that its own figure is a floor: "this data represents just the visible AI traffic, the minimum amount of AI visitors a website receives," because some AI traffic pulls through labelled as Direct. Treat every traffic number in this piece the same way.

Does that traffic actually convert?

Volume answers one question; conversion answers the one that decides whether AI visibility is worth chasing. Semrush studied conversion rate across more than 500 high-value digital marketing and SEO topics and found the average AI-search visitor, tracked to a source like ChatGPT, is 4.4 times as valuable as the average visitor from traditional organic search, measured on conversion rate.

The mechanism behind it is straightforward. An AI answer has usually already compared the visitor's options and half-decided for them before they click, in a way a search results page rarely does. Someone who typed a plumber's postcode into Google is still choosing; someone who clicked a name ChatGPT recommended has already been sold most of the way. Engagement data backs this up from a different angle: across 304 client properties measured over two years, ChatGPT-referred visits engaged at 63.42%, against 61.64% for organic search from the same properties, in Siege Media's engagement study. Higher engagement and a 4.4x conversion multiple point in the same direction: this is not lower-quality traffic padding out a vanity metric.

QBiz's own research shows how little of that ground is currently claimed. In the AI Visibility Gap US study, QBiz audited 191 US local service business websites, plumbers, attorneys, dental practices and accountants, against ten binary readiness signals. All 191 were scored on a single collection date in August 2026. Among the 179 sites that fetched and scored, technical delivery was close to universal: 97.2% rendered correctly on mobile, 96.1% were crawlable, and 93.9% carried adequate internal linking. Structural signals sat lower, with service definitions on 59.8% and schema markup on 50.8%. Answer-oriented content was rarest of all: 35.2% had an llms.txt file, 20.1% explained their own process, and 18.4% carried FAQ content. Among the sites in that sample that an audit could read at all, almost all were technically readable and few offered a ready-made answer to quote. That content gap is what AI visibility work closes. The study scored website readiness signals, not AI citations or leads.

Why the traffic is shifting in the first place

Part of why this matters now: Google's own organic results are sending fewer clicks whenever an AI summary appears. Pew Research tracked 900 US adults across nearly 69,000 Google searches in March 2025 and found that when an AI summary appeared, users clicked a traditional result in just 8% of visits, against 15% when no summary showed, and ended the session without clicking anything more often too (26% versus 16%). Whether the clicks that go missing from organic search reappear inside an AI platform's own answer is a separate question, covered next.

Why AI referrals are hard to attribute

AI clicks often arrive with no referrer, so Google Analytics files them as Direct alongside someone who typed the address from memory; a genuine AI-driven visit and an old customer typing your URL look identical in the data. That means no lead count from AI visibility, on this page or anywhere else, can currently be complete. The mechanics of measuring it as closely as today's tools allow, GA4's AI Assistants channel, Search Console's generative-AI reports, and their limits, are covered in full in How to Measure AI-Search ROI.

Where the answer is genuinely "it depends"

A single yes-or-no answer would be false precision. Four things move the real answer for a specific business:

Vertical and query type. AI platforms answer "who should I hire for X near me" and "what's the best way to do Y" more confidently in some sectors than others, largely based on how much clearly structured, citable content exists for that trade or profession online. A business in a sparsely documented niche may see almost no AI-referred traffic yet, because there's little for a model to cite.

Current traffic volume. A business getting ten AI-referred visits a month cannot draw a statistically meaningful conclusion about lead quality from that sample, whatever the visits convert at.

How much is already hiding in Direct. Every business's own Direct traffic bucket almost certainly contains some AI-driven visits it currently has no way to separate out. A business with unusually high or rising Direct traffic and no obvious explanation (no new offline campaign, no press mention) has a specific, checkable reason to look closer, covered step by step in How to Check If AI Recommends Your Business.

Time horizon. The growth rate in the Similarweb data (357% year on year) describes a channel still early in its adoption curve. A business judging AI visibility purely on this month's traffic numbers is measuring a fast-moving target with a slow-moving yardstick.

None of these four factors can currently be reduced to a single confidence score. A business genuinely doesn't know its own answer until it looks at its own data against these four questions, which is a fair, if unsatisfying, thing to say. It's also worth flagging that the evidence base above is predominantly US-sourced (Similarweb, Ahrefs, Siege Media, Pew and Semrush all study US or global-skewed samples, and QBiz's AI Visibility Gap study is US-only), so read every figure on this page as the best available proxy for a UK business, not a measured UK number.

What can be measured today, and what can't

Measurable now: whether AI platforms can read and parse a site's content at all (a technical, checkable fact). Whether a business is being mentioned or recommended by name when someone asks an AI tool a relevant question, through manual prompt testing or a monitoring tool, covered in How to Check If AI Recommends Your Business. Engagement metrics (time on page, pages per session, bounce) for the traffic that is visibly tagged as AI-referred.

Not reliably measurable today: a precise revenue or lead figure attributable specifically to AI visibility work, separated from organic SEO, brand awareness, or word of mouth, because these channels increasingly overlap and reinforce each other rather than operating in clean, separate lanes. Real-time, prompt-by-prompt monitoring of every question a potential customer might ask an AI tool, at the level of granularity a full-funnel lead count would need.

How to think about the investment given all of that

Given a genuinely partial picture, the sound approach isn't to wait for perfect measurement before doing anything. Two practical positions hold up:

Weight the decision by vertical and current visibility rather than an industry-wide average. A business already well-optimised for AI platforms and getting cited regularly is in a different position from one that has never checked whether AI tools mention it at all; the second group has more obvious upside because the starting point is close to zero.

Separate the technical foundation from the traffic question. Structuring a site so AI platforms can read and cite it correctly, valid markup, clear answer-first content, accurate business information, is worth doing regardless of how the attribution debate settles, because the same foundations also support ordinary SEO. The traffic-measurement gap is a reason to stay realistic about reporting, not a reason to skip the underlying work.

FAQ

Can any agency currently prove exactly how many leads AI visibility drives?

No agency can currently isolate a precise lead count attributable specifically to AI visibility, because a meaningful share of AI-referred traffic arrives in analytics platforms as unlabelled Direct traffic.

Is AI referral traffic worse quality than traffic from Google search?

The evidence points the other way. Semrush found the average AI-search visitor is 4.4 times as valuable as an average organic-search visitor, measured on conversion rate, and a separate study of 304 client properties found ChatGPT-referred traffic engaging at 63.42%, against 61.64% for organic search. Neither is a universal guarantee, but both cut against the assumption that AI-referred visitors are lower-intent than search visitors.

Why does AI traffic show up as "Direct" in my analytics?

Analytics platforms classify a visit as Direct when they detect no referrer and no campaign tag. Many AI assistants pass little or no referrer information when a user clicks a link inside an answer, so a genuine AI-driven visit frequently looks identical, in the data, to someone typing a web address from memory.

Should a small business invest in AI visibility now or wait for better measurement tools?

That depends on the four factors covered above: vertical, current traffic volume, how much unexplained Direct traffic already exists, and time horizon. Waiting for perfect measurement means waiting indefinitely, since full attribution isn't close. Getting the technical and content foundations right carries a return regardless, which is a reasonable way to start without depending on measurement that doesn't exist yet.

Short version, restated

The traffic is real, growing fast off a small base, and it converts well once it arrives. Judge the investment by your own vertical, your own current visibility, and your own Direct-traffic pattern rather than a single industry-wide number.

Want to know where your own leads are actually coming from before you spend more on this? Check whether AI platforms can find and cite your business →