Independent Analysis: AI, SaaS, and Revenue Growth Strategy

the number Anil Chakravarthy led with at the Adobe Summit keynote this morning was 269 percent. That is how much AI-originated traffic to US retail websites grew in March 2026, year over year, according to Adobe Analytics data covering more than a trillion visits. The number was not the headline of any press release. It was the foundation underneath every product announcement Adobe made today. Figures cited throughout are Adobe-sourced and unaudited by any third party.

Most coverage is going to treat 269 percent as a growth story. It is not. It is a reversal story, and the reversal is what should reset the conversation inside every chief marketing officer's staff meeting this quarter.

Twelve months ago, AI shoppers were the worst customers on your website

In March 2025, AI-originated traffic to US retail sites converted 38 percent worse than traffic from paid search, email, and other traditional channels. Bots browsed. Humans bought. Every retail analytics dashboard in the country told the same story: invest in the channels that produced the purchases, treat AI traffic as curiosity.

In March 2026, AI traffic converted 42 percent better than non-AI traffic. That is a new record, according to Adobe. Revenue per visit from AI referrals was 37 percent higher. Engagement was 12 percent higher. Time on site was 48 percent longer. Pages per visit were 13 percent higher.

One year ago, human traffic was worth 128 percent more than AI traffic. Today, the AI shopper is worth more than the human one by every retail metric that matters.

That is not a trend line. That is a phase change.

Why the reversal is structural, not seasonal

The instinctive pushback is that retail had a strong Q1, AI hype inflated traffic, and the conversion numbers will normalize. That would be a comfortable read. It is probably wrong, for two reasons.

First, the holiday season preview was even sharper. AI traffic grew 693 percent year over year in November and December 2025. The first quarter of 2026 overall was up 393 percent. March's 269 percent is a deceleration from holiday peaks, not an acceleration driven by one month's demand spike. The baseline is resetting upward even as growth rates compress. That is what real adoption curves look like.

Second, the conversion quality improvement is consistent with how AI-assisted shopping actually works. An AI shopper arrives on a product page already pre-qualified. The model has already considered alternatives, weighed reviews, and matched the product to the stated need. The human clicking through from a paid search ad is still comparing. The human clicking through from an email campaign is responding to a discount, not a decision. The AI-referred visitor is further down the funnel before they arrive, and the funnel is shorter.

If that analysis holds, the conversion advantage does not normalize. It compounds.

The invisibility problem Adobe is selling against

Adobe said on stage that 80 percent of businesses have significant gaps in how their brands show up in AI platforms. The company also reported that roughly one quarter of homepage content on US retail sites is not optimized for large language models to read, and category page content fares similarly.

Translate that into a chief marketing officer's language. You are underinvested in the channel that is now producing your most valuable shoppers, and you cannot see the channel clearly because your analytics were built for a web where Google was the intermediary. Adobe knows this because Adobe Analytics covers more than a trillion visits to US retail sites, which is a larger dataset than anyone else has on the subject. The company is selling the observation and the product to fix it in the same breath.

That is what Brand Visibility, LLM Optimizer, Brand Concierge, and the pending Semrush acquisition are for. They are a product stack aimed at the gap between the traffic retailers are already receiving and the optimization their websites have not yet done for that traffic to find them correctly.

The marketing measurement problem just got sharper

Most enterprise analytics stacks were built to attribute conversion to paid search, email, display, social, and direct traffic. AI-originated traffic is arriving through referrers that do not fit those categories cleanly. A visit that began in a ChatGPT conversation and ended in a purchase gets attributed as direct or referral depending on how the AI platform passes through headers, which varies by provider and by month.

That means your 2025 marketing mix model, the one that informed your 2026 budget allocation, almost certainly misattributed the highest-performing channel you have. Some percentage of what your attribution system credited to email, paid search, or direct was actually AI-assisted discovery. You overpaid the channels you could see and underfunded the channel you could not.

The marketing mix model for fiscal 2027 needs a new line item. Building it requires the kind of log-level analytics and referrer-parsing infrastructure most teams have not yet invested in. Adobe is one of the vendors who can sell that infrastructure. So is Google. So are a handful of specialists. The question is whether you pick a vendor before your next budget cycle or after.

What this does not prove

Two caveats worth sitting with.

The data is retail-specific. Adobe's trillion-visit dataset covers US retail sites. Financial services, healthcare, business-to-business software, and education may be on different curves. The conversion reversal may not repeat cleanly in categories where the purchase decision does not lend itself to AI pre-qualification the same way a product comparison does.

The figures are Adobe's alone. No independent research firm has validated the 269 percent, 42 percent, or 37 percent numbers. Adobe has commercial incentive to make AI traffic look important, because Adobe sells the optimization stack. That does not mean the numbers are wrong. It means they should be treated as directional evidence from an interested party until independent measurement catches up.

I walked into the Summit keynote this morning ready to write about agents. Instead I walked out thinking about a measurement reversal most chief marketing officers do not yet have a line item for.

The traffic you are receiving is changing faster than the analytics you are using to measure it. The visitors your website was designed for are not the ones arriving anymore, and the ones arriving are spending more money than the ones you designed for.

That is not an Adobe product story.

That is the story the Adobe product story is sitting on top of.