AI Is Rewriting Search. Most Online Stores Are Still Failing the Readiness Test
ResourcesNews

AI Is Rewriting Search. Most Online Stores Are Still Failing the Readiness Test

AI referral traffic to retail sites is set to jump 130% this holiday season, but most stores still fail basic AI search readiness checks.

By VTEXNov 17, 20265 min read
AI search readinessAI Commerce Rankingsstructured product dataagentic commerce protocolsAI referral traffic

AI referral traffic is no longer a rounding error

AI search readiness is becoming a measurable gap, not an abstract concern, as AI-referred traffic to U.S. retail sites is forecast to grow 130% year over year during the 2026 holiday season, with the sharpest spikes on Thanksgiving (159%), Black Friday (95%) and Cyber Monday (88%), according to Adobe's holiday forecast reported by Digital Commerce 360 [1].

Since Adobe started tracking AI referrals in October 2024, that traffic has climbed 1,219% as of July 2026, separate from the holiday forecast [1]. Adobe based both figures on more than 1 trillion visits to U.S. retail sites, 100 million SKUs and 18 product categories [1].

The behavioral shift behind the numbers matters as much as the growth rate. Among shoppers who used AI for online shopping, 77% said they felt more confident in their purchase and 69% said they were less likely to return the item, per a complementary Adobe survey of 5,000 U.S. consumers [1]. Confidence and lower returns change the economics of a channel that was marginal two years ago.

Most retail catalogs still fail the AI search readiness test

Across the Top 1,000 US online retailers, the average AI shopping readiness score is 41.9 out of 100, and only 20 retailers score above 60, according to the AI Commerce Rankings built by Digital Commerce 360 and ReFiBuy [3]. No retailer in the index scores above 71.8.

The index scores four signals that reflect what an AI shopping agent actually encounters when it reaches a catalog: bot friendliness, AI source traffic, diversity of AI sources, and 90-day momentum [3]. A low score does not mean a retailer has no AI strategy in progress; it means agents cannot yet see the results of that work [3].

Readiness does not track with sales scale. Online Labels, a packaging and labels supplier ranked No. 814 by online sales, leads the index with a readiness score of 71.8, ahead of household names many times its size [3]. ChatGPT alone accounts for more than 80% of AI-referred retail traffic in the first edition, which is why source diversity counts as a scored signal rather than a footnote [3].

CategoryAvg. readiness scoreWhat stands out
Office supplies47.4 [3]Highest score of any category despite its small size
Jewelry45.0 [3]Most diversified answer-engine visibility in the index
Hardware & home improvement44.8 [3]Led by niche specialists, not household names
Mass merchant38.5 [3]Sits 179 places lower on AI readiness than on sales rank
Food & beverage37.2 [3]Lowest score and lowest AI traffic penetration in the index

AI traffic converts better, but it is still a fraction of total volume

AI-referred shoppers added items to cart at a 32% higher rate and generated 43% more revenue per visit than non-AI traffic, a complete inversion of the pattern Adobe measured 18 months earlier, when non-AI traffic was worth 128% more per visit [2].

“It also behaves as a discovery channel through AI search and a referral channel through the AI chatbots themselves.”
Rebecca Wettemann, CEO and principal analyst, Valoir [2]

The scale of the shift should not be mistaken for the scale of the channel itself. Total online holiday spend is forecast to grow 6.7% in 2026, reaching $275.1 billion, while AI referral traffic grows from a much smaller base [1]. The gap between those two growth rates is the real story: a small channel is becoming the highest-value channel a retailer has, even before it becomes the largest one.

That framing matters operationally. A discovery channel and a referral channel demand different fixes: discovery depends on whether an agent can find and parse a product at all, while referral conversion depends on what happens once a shopper lands on the site.

What AI shopping agents actually check before they recommend a product

AI agents score a page's machine-readability before deciding whether to recommend it, and most retail categories still fall well short of full readability. Adobe's AI Content Visibility Checker found cosmetics pages were 63% AI-readable and electronics pages 56%, while grocery (48%) and furniture and home (47%) trailed, facing what Adobe called structural challenges in page design that suppress AI citation [4].

The gap shows up unevenly even within a single category. Cosmetics blog and news pages were 78% AI-readable, but grocery's strongest page type, product detail pages, reached only 70% [4]. A retailer can be ahead on one page type and well behind on another, which is why a single category-wide score can hide where the actual fix needs to happen.

The AI Commerce Rankings' bot-friendliness signal captures a related but distinct gap: whether agents can transact with a catalog at all, not just read it [3]. That includes verified support for commerce protocols like Universal Cart Protocol and Agentic Commerce Protocol, categories where the rankings found most retailers, including some of the largest, still have no verified status [3].

Four checks to run before your next catalog review

A readiness audit does not require new tooling before it requires discipline about what already exists. The following sequence mirrors the signals both Adobe and the AI Commerce Rankings use to score retailers, adapted into steps a commerce team can run internally.

Universal standards for AI product feeds do not yet exist, but a retailer already maintaining a clean feed for traditional shopping engines is closer to ready than one starting from scratch [5].

  1. 1Audit product feed completeness the same way you would for Google Shopping: price, availability, GTIN or MPN, and return policy fields filled for every active SKU, not just top sellers [5].
  2. 2Verify protocol support for agent checkout, specifically whether your platform has confirmed status for Universal Cart Protocol or Agentic Commerce Protocol [3].
  3. 3Diversify AI source visibility instead of optimizing for one assistant. A retailer visible only on the dominant one inherits that platform's risk [3].
  4. 4Track AI referral traffic as its own line in analytics, not folded into "other," so conversion and revenue-per-visit differences are visible before the next peak season [5].

Is your e-commerce ready for agent-driven shopping, or just lucky with traffic?

Readiness for agent-driven shopping means a catalog agents can find, read and transact against, not just a site that happens to receive AI clicks today. The distinction matters because traffic gains from the current wave of consumer AI assistants do not guarantee standing with the agent-to-agent commerce now scaling on the B2B side.

Gartner projects that AI agents will intermediate 90% of B2B purchases, representing $15 trillion in spend, by 2028 [7]. On the consumer side, eMarketer projects AI platforms will account for $20.9 billion in U.S. retail spending in 2026, nearly quadrupling 2025 levels [7]. Both numbers describe the same underlying shift at different speeds.

Retailers are already splitting on strategy in response. Amazon blocked Meta's Muse AI shopping agent to protect its own data and customer relationships, while Shopify welcomed it onto its merchants' stores, reflecting two different bets on where the customer relationship will sit once agents mediate more of the buying journey [6].

The readiness gap is the opening, not the verdict. Most of the Top 1,000 has room to move, and catalog discipline closes the gap faster than sales volume does [3].

Sources

  1. [1]Digital Commerce 360, Adobe projection: 2026 online holiday sales to reach $275.1 billion — digitalcommerce360.com
  2. [2]Retail Dive, Consumer AI use a boon for the holidays as AI traffic doubles — retaildive.com
  3. [3]Retail TouchPoints, New Rankings Rate AI Retail Readiness Across 1000 Companies — retailtouchpoints.com
  4. [4]Digital Commerce 360, Adobe: AI-referred traffic to retail sites doubles in a year — digitalcommerce360.com
  5. [5]Practical Ecommerce, Next Steps for AI Shopping — practicalecommerce.com
  6. [6]Retail TouchPoints, Are Shoppers Ready for Muse AI Agentic Commerce? — retailtouchpoints.com
  7. [7]Forbes, Agentic Consumers Are Creating A Trillion-Dollar Market Shift — councils.forbes.com