Adobe Just Gave the Discovery-Layer Thesis a Number: AI Traffic to US Retail Sites Up 393%, Converting 42% Better than Humans
Richard Lee
April 22, 2026 · 4 min read
On April 16, Adobe Digital Insights director Vivek Pandya published what may be 2026's most important retail data point. Adobe Analytics, tracking over one trillion visits to US retail sites, found AI-referred traffic up 269% year over year in March, up 393% for Q1, and, headline figure, AI now converts 42% better than non-AI traffic. That's a 100-point swing from March 2025, when AI converted 38% worse than paid search and email. Revenue per visit from AI runs 37% higher. Six weeks after OpenAI retreated from in-chat Instant Checkout, the discovery layer it kept is now US retail's most valuable traffic source.
Yesterday I wrote about Mubboo's discovery-layer positioning. Today Adobe put a number on it. I want to be specific about what Adobe confirms, where it doesn't generalize, and why the 34% of product page content LLMs can't read matters.
What Adobe measured
Pandya's report draws on Adobe Analytics' tracking of over one trillion visits to US retail sites. AI-referred traffic grew 269% year over year in March 2026 and 393% across Q1. Holiday 2025 AI traffic had grown 693% between November and December. The conversion picture flipped in twelve months: in March 2025, AI converted 38% worse than paid search and email; in March 2026, it converts 42% better. Revenue per visit from AI runs 37% higher than non-AI. A year ago, human traffic was worth 128% more per visit. AI visitors also spend 48% more time on site, browse 13% more pages, and post a 12% higher engagement rate.
Why this happened — the trust signal beneath the traffic
The traffic surge sits on top of a trust shift. In Adobe's surveyed pool, 66% of consumers believe AI tools provide accurate results. 39% have used AI for online shopping, and 85% of that group said it improved their experience. Pandya's framing: "Rising consumer trust has played a factor."
Pair that with yesterday's OpenAI Instant Checkout retreat coverage, where 8% of US adult ChatGPT users tried in-chat checkout in its first month, and the trust conversation splits into two problems. Discovery trust asks whether the recommendation is honest and accurate; Adobe measured 66%, with traffic up 393%. Transaction trust asks whether checkout, payment, delivery, and returns hold together; OpenAI measured 8% adoption and pulled back. Quad's 75% trust cliff and Adobe's 66% accuracy trust live in the same consumers, on different layers.
The 34% problem — why it matters for editorial layers
Adobe introduced a new metric in the report: the AI Content Visibility Checker. Average homepage visibility to LLMs sits at 75%, with the highest-scoring retailers reaching 82.5% and the lowest at 54.2%. Average product-page visibility sits at 66%, meaning 34% of product page content is invisible to large language models. Product pages are where purchase decisions happen, which makes that gap the meaningful one.
Pandya's read: retailers carry thousands of SKUs, and much of that content sits invisible to LLMs today. AI is shaping the most valuable traffic in US retail but can see only about two-thirds of what sits on the pages that traffic lands on. Editorial written explicitly to be cite-able (structured data, clear attribution, no CMS-hidden text) fills the part LLMs can't read.
What this means this week, specifically
The data is six days old and the implications are still landing. Read alongside Phocuswright's 6% scaling agentic AI, Stanford AI Index 2026's single-digit agent department adoption, and OpenAI's March 5 retreat, Adobe's numbers tell the other side of the story. Agentic AI for checkout is scaling slowly under trust and operational friction. AI for discovery is scaling fast, outperforming paid search and human traffic per visit.
These aren't contradictions. They are two layers of the same system. The scaling layer, discovery, rewards specific, cite-able content. The other, transaction, is being handled at the merchant. The editorial content sitting between the two is the discovery layer.
Mubboo's take
I wrote Mubboo's discovery-layer positioning into the Blog yesterday. Fewer than 24 hours later, Adobe's April 16 report is doing the validating. AI traffic is now the highest-converting source US retailers have, and revenue per AI visit runs 37% higher than anywhere else. Yet 34% of product page content is invisible to the LLMs driving that traffic. If a retailer's page holds purchase-relevant context an LLM can't read, the LLM reaches for editorial it can. That's the Mubboo position, stated precisely. We aren't fighting AI for discovery attention. We feed it daily, on mubboo.com/shopping. Adobe's data tells us the 34% gap isn't a retailer problem. It's our opportunity.
One number worth ending on: per Adobe, the highest-scoring retailers hit 82.5% LLM visibility on their homepages, the lowest 54.2%. For editorial like us, the work is to stay in the 90s. The difference is between being cited and being invisible.
Richard Lee
Founder
Richard is the founder of Mubboo, building an AI-powered platform that helps everyday consumers navigate shopping, travel, finance, and local life across multiple countries.