Adobe: Prime Day genAI shopping traffic up 103% YoY, $26.3B in play

What Adobe published

On June 23, 2026, Adobe Digital Insights published its Prime Day 2026 forecast: U.S. online spend across the four-day June 23 to 26 window is projected at a record $26.3 billion, a 9% lift versus 2025. The headline that matters more to sellers is buried underneath: traffic to U.S. retail sites originating from generative-AI sources is tracking up 103% versus June 2025 (WWD / Sourcing Journal coverage of the Adobe forecast).

Digital Commerce 360 framed the same forecast against last year's BFCM weekend and noted the projected take is higher than 2025 Black Friday and Cyber Monday combined (Digital Commerce 360). The center of gravity for upper-funnel discovery has shifted noticeably in twelve months.

The headline numbers

Why AI-surfaced traffic behaves differently

Shoppers landing on a product detail page from Rufus inside Amazon, ChatGPT in a browser tab, or Perplexity in an answer card did not click through a Sponsored Products slot to get there. The agent summarized the page first and decided to send the user. What the agent read to make that decision is structured: the attribute table, A+ structured fields, review aggregates, answered questions, and the price-per-unit rollup.

The listing that ranks well on keyword search but has thin attribute coverage or unanswered top questions tends to be skipped by an agent, even if it wins the Sponsored slot. The lever that moves AI-surfaced sessions is not bid management. It is the quality of the structured surface area on the detail page.

What to audit on hero ASINs before Day 3

• Attribute table coverage. Fill every brand-registry attribute, not just the required fields. Agents read the optional ones too, and a complete table is a strong relevance signal.
• A+ structured modules. Comparison charts and feature tables in A+ content are machine-readable. Free-form lifestyle banners are not. Both have a role, but the structured modules carry the agent read.
• Q&A density on recurring objections. Every repeated buyer question that has gone unanswered for more than a week is a gap an agent will flag. One answered question per recurring objection is the floor.
• Price-per-unit and pack-size hygiene. Agents lean on price-per-unit to compare across listings. A missing or wrong rollup makes the listing invisible to that comparison.
• Review recency. Aggregate star ratings matter less than whether the recent reviews mention the attributes the page claims. A mismatch reads as a trust penalty in the agent summary.

How Nova helps

• Day-to-Day Analytics tracks session-to-unit conversion on hero ASINs daily so the AI-surfaced traffic uplift can be measured, not assumed.
• PPC Analytics separates ad-driven sessions from organic and direct so the share of traffic that does not pay an ad fee is visible.
• BSR Tracker trends ASIN rank by marketplace so listings improved during the event can be re-evaluated against rank movement, not just sales.