For a long time, the AI traffic —visits that land on your website after someone asks ChatGPT, Gemini, Perplexity or Copilot a question—used to be a minor item in analytics reports. Something anecdotal, easy to overlook. By 2026, that picture has changed completely: as well as growing at a striking rate, this traffic converts better than almost any other source. And we’re not just saying this on a hunch, but because the data is right there on the table.
According to Adobe Analytics, whose findings were reported by Digital Commerce 360, 17 June 2026, visits to retail websites from AI sources converted 54 % better than those from non-AI sources during May 2026. The analysis is based on more than one trillion (one million million) visits to US e-commerce sites. At Vandelay, we believe it’s worth taking a moment to understand what lies behind that figure, as it shifts a few priorities.
What exactly does the Adobe report say?
Short answer: traffic from AI assistants is no longer marginal in terms of either volume or quality. Here are the verifiable figures (Adobe Analytics, May 2026):
- +138 % year-on-year. Referral traffic from IA to US retail websites grew by 138 % compared with May 2025.
- +1,324 % since October 2024. More than 14 times, since Adobe began tracking these figures.
- Conversion to a 54 % superior compared with non-AI traffic. This marks a reversal from the previous year, when that same channel accounted for nearly half of conversions.
- +15 % engagement. Users are more engaged; this has been the biggest difference compared with non-AI sources since October 2024.
- 53 % more time on the site and 23 % more pages views per visit.
An important caveat to avoid over-interpretation: these figures relate to the US retail sector and websites that use Adobe Analytics. They provide a very strong indication of the direction things are heading, but they do not give an exact picture of the Spanish market. That said, the trend is consistent with what we have been observing in European accounts.
Why does AI traffic convert better?
The sensible explanation isn’t that AI is magic, but that it changes the moment a person lands on your website. Previously, someone would open Google, compare ten tabs and land on your page right in the middle of their research. Now, much of that research takes place within the chat: the assistant summarises options, answers questions and rules out alternatives. By the time they finally click through to your website, they’re much further along in their decision-making process.
Adobe had already described this back in 2025 in its analysis on the rise of AI-driven referral traffic: users are «further along in the consideration process». In other words, they are less likely to be simply browsing and more likely to be buyers. This ties in with the 53 % increase in time spent on the page and the higher number of page views: these aren’t people who have got lost; they are people comparing the specific product that the AI suggested to them.
There is a second factor. AI tends to perform better when it comes to complex or expensive purchases. In the same analysis, Adobe noted that AI-driven traffic performed particularly well in research-intensive categories, such as consumer electronics and jewellery, whilst the advantage was less pronounced in fashion and food. Put another way: the more the customer thinks about it, the more it pays to appear in the assistant’s response.
Not all sectors start on an equal footing: the readability gap
Here comes the awkward bit. For AI to make recommendations, it first needs to be able to read and understand your website. Adobe measures this using its AI Content Visibility Checker, which assigns each page a readability score for language models. And the results from May 2026 show notable differences by sector:
| Sector | Readability for AI (May 2026) |
|---|---|
| Cosmetics | 63 % |
| Electronics | 56 % |
| Sport and fashion | 51 % |
| Food | 48 % |
| Home and furniture | 47 % |
The practical implication: even the best-performing sectors leave between 30 and 40 % of their high-value content unrecognised by AI, according to Adobe. Pages with clear specifications, lists of ingredients, tutorials, guides and customer service pages are exactly the kind of content that models «read» best. If your website hides that information within images, PDFs or scripts, you’re leaving it out of the conversation.
What can you do this week to attract AI traffic?
There’s no need to reinvent your strategy. These are the steps we take at Vandelay when a client wants to make the most of this channel:
- Measure first. In your analytics, segment the visits coming from AI domains (such as chatgpt.com, perplexity.ai, gemini.google.com or copilot.microsoft.com, amongst others) and compare their conversion rates with the rest. Without that data, the rest is just opinion.
- Make your content machine-readable. Actual text in HTML (not embedded in images), clear headings and product pages with specifications and direct answers.
- Answer questions, not just keywords. AI quotes content that answers specific questions. A well-written Q&A section is more effective than a generic paragraph.
- It reinforces signals of trust. Reviews, contact details, authorship and verifiable sources help ensure that a model is treated as a reliable recommendation.
- Pay attention to the whole document, not just the cover page. Many of these visits go straight to product or service pages. If the homepage looks great but the product or service page is poor, you’ll lose conversions where it really matters.
If this sounds like SEO to you, it’s because they share the same roots. The difference is that here the aim isn’t just to appear on a list of links, but to be the source that the AI chooses to use to provide an answer. It’s what we in the industry call GEO (optimisation for generative engines), and now is a good time to take it seriously.
Frequently Asked Questions about AI traffic
What is AI traffic?
This refers to the total number of visits to your website from artificial intelligence assistants — such as ChatGPT, Gemini, Perplexity, Copilot and similar services — usually after a user has made a query and the assistant links to or mentions your page.
Does AI traffic really convert better than Google’s?
According to Adobe Analytics data from May 2026, AI traffic converted 54 % better than non-AI traffic on US retail websites. This is not a direct «AI versus Google» comparison across all sectors, and results vary by category, but the trend towards improvement is clear and sustained.
Is this suitable for an SME in Spain?
The data is from the US, so it’s best to treat it as a trend rather than a local figure. That said, the use of AI assistants for searching and shopping is also growing in Europe, and the same best practices for readable and useful content apply here too. It’s advisable to measure your own AI-related traffic before drawing any conclusions.
Do I need expensive tools to get started?
No. Start by segmenting the traffic in your current analytics and checking that your key pages contain genuine, clear and well-structured text. From there, you can set your priorities.
The big picture is simple: more and more people are making up their minds in a chat before they even visit your website, and those who arrive from there are more likely to make a purchase. Ignoring this won’t stop the trend; it just means your competitors will be the ones appearing in the reply. At Vandelay, we’re helping brands in Barcelona and across the rest of Spain to measure this channel and optimise their websites so that AI recommends them. If you want to know how much AI-driven traffic you’re already receiving — and what it’s doing with it — that’s a good place to start the conversation.
English (UK) 
Español