Three research firms dropped new numbers on AI shopping this month, and together they say something most Shopify merchants have not caught up to yet. AI is no longer sending you a trickle of curious browsers. It is sending traffic that converts better than the rest of your channels, from shoppers who are willing to hand a budget to an agent, as long as they can send the box back. Here are the four numbers worth planning against, plus the five checks in your Shopify admin that decide whether any of it reaches your store.
1. 59% of US shoppers say AI showed them a brand they did not know
RTB House surveyed 1,840 people across the US, UK, France, and Japan in June and July 2026 with Cint. 59% of US respondents said AI is good at surfacing brands they had not previously heard of, against 52% outside the US.
The age split is the part nobody predicts. Gen X came in at 61% and baby boomers at 62%, the two most receptive groups in the study. This is not a young shopper story. It is the demographic with the highest average order value in most parts categories telling you it now meets new brands inside a chat window.
Discovery is moving into the model, not the results page. If you have watched organic sessions soften over the last eighteen months, that is the other half of the same trend we covered in how AI Overviews and LLMs are changing Shopify SEO.
2. 42% of US millennials will give an agent $250, if they can return it
Same study. 42% of US millennials said they are comfortable giving an AI agent a budget of up to $250 to buy on their behalf when a seven day return window is guaranteed. Non US millennials sat at 27%.
Now take the guarantee away. Comfort drops to 34% in the US and 20% elsewhere. Eight points of agent checkout willingness live inside your return policy page. Approval before checkout was the safeguard shoppers asked for most, everywhere.
One more line from the same data that cuts the other way: 42% of US respondents said AI tools make purchase decisions take longer, rising to 48% for Gen Z. AI compresses discovery and stretches deliberation. More comparison, more touches, and more weight on whatever text a model can quote back about your product.
3. The 4,700% number is a year old. The 2026 numbers are more useful.
Adobe Analytics measured generative AI referral traffic to US retail sites up as much as 4,700% year over year in July 2025. That figure is in every agency deck, and it is a year stale. Here is what Adobe has published since:
- Holiday 2025 (November to December): AI traffic up 693% year over year
- Q1 2026: up 393% year over year
- March 2026: up 269% year over year
Growth rates are cooling because the base is finally real. The quality is what changed. In March 2025, AI sourced traffic converted 38% worse than non AI traffic. In March 2026 it converted 42% better, a record. AI referred visitors also showed a 12% higher engagement rate, stayed 48% longer, and viewed 13% more pages per visit.
Read that as a shift in intent. Early AI traffic was people poking at a new toy. Current AI traffic is people who already did their research inside the model and arrived to buy.
4. McKinsey: $3 to $5 trillion globally by 2030, up to $1 trillion in the US
McKinsey estimates AI agents could mediate $3 trillion to $5 trillion of global consumer commerce by 2030, with the US portion running up to $1 trillion. Forecasts that far out are directional, not a plan. What is worth reading is the merchant guidance underneath the number, because none of it is about advertising.
McKinsey tells retailers to make catalogs machine readable or accept invisibility to agents regardless of brand strength, to move to API first merchandising with inventory, pricing, shipping promises, and promotions exposed as clean structured data, and to treat reversibility as a competitive advantage: easy cancellation, refunds, and human override.
Reversibility is the same finding as the RTB House return window number, arriving from a completely different direction. That is usually a sign the finding is real.
The number nobody quoted: product pages are your least readable pages
Buried in Adobe's 2026 data is a machine readability score for retail page types, meaning how much of the page an LLM can actually parse. The results are backwards from what you would guess:
- Returns and exchanges pages: 82%
- Contact pages: 81%
- FAQ pages: 80%
- Customer service and help: 79%
- Loyalty and membership: 78%
- Homepages: 75%
- Category pages: 74%
- Product pages: 66%, the worst score of any page type
Roughly a third of the content on a typical retail product page is invisible to a model. The product page is where compatibility, specs, materials, and fitment live. For an automotive parts store, that is the entire sale. Your policy pages parse fine. The page that decides the purchase does not.
We wrote about why this happens in parts catalogs specifically in fitment data is your AEO strategy for auto parts.
Five checks in your Shopify admin that take an afternoon
Confirm Agentic Storefronts is actually on
Settings, then Sales channels, then Agentic Storefronts. Shopify activates it automatically for eligible merchants and it syndicates products to ChatGPT, Microsoft Copilot, AI Mode in Google Search, and the Gemini app with no app to install. Automatic is not the same as verified. Check the channel status and, more importantly, how many of your products it actually publishes.
Pull the Sessions by referrer report
Shopify Analytics, Reports, Sessions by referrer. Look for chatgpt.com, perplexity.ai, gemini.google.com, claude.ai, and copilot.microsoft.com. Then go a level deeper: filter orders by each source and compare average order value and return rate against your site average. Adobe's national numbers are a benchmark, not your number. A parts store with thin fitment text often sees the opposite pattern, with higher AOV and a much higher return rate.
Get fitment detail into the product description
If your store has year, make, and model functionality, that data almost certainly lives in Shopify metafields feeding a front end widget. An agent reads the retrievable text on the page. Put the covered vehicles, MPN, GTIN, real stock state, and price into the product description and structured data, not only into the widget. This is the single highest leverage fix on the list, and it is also the one that quietly drives returns, which we broke down in Shopify auto parts returns are a fitment data problem.
Check robots.txt.liquid for AI crawler blocks
Plenty of stores added crawler blocks in 2023 and 2024 and never revisited them. If GPTBot, ClaudeBot, or PerplexityBot are disallowed, you are invisible to the traffic that now converts 42% better than everything else.
# In robots.txt.liquid, a block looks like this. Find it, then remove it. User-agent: GPTBot Disallow: / User-agent: ClaudeBot Disallow: / User-agent: PerplexityBot Disallow: / # Also check the live file at yourstore.com/robots.txt # A store with no custom robots.txt.liquid uses Shopify's default, which allows them.
Rewrite your return policy so a model can parse it
Given the 42% versus 34% split, this page is a conversion asset, not legal boilerplate. State the return window in days, the restocking fee as a number, who pays return shipping, core charge handling, and any exceptions for electrical or special order parts. Plain text on the page. Not a PDF, not an image, not a link to a portal that requires a login.
What we would check first
If you only do one thing this week, pull Sessions by referrer and compare the return rate of AI sourced orders against your site average. That single comparison tells you whether models are describing your products accurately or guessing. Everything else on the list is a fix for whatever that number shows you.
Ambaum runs AI readiness audits for Shopify Plus automotive parts merchants. We look at what models actually see when they read your catalog, your policies, and your product pages, then hand you the gaps in priority order. Book an agentic commerce audit and find out where you stand before the holiday season does it for you.
Read the full transcript
New AI eCommerce reports landed this month. Here are the four numbers Shopify merchants should be planning against.
Nearly 60% of US consumers say AI platforms showed them brands they did not already know. Discovery is moving into the model, not the results page.
42% of American millennials would let an AI agent buy for them within a $250 budget, as long as they can return it inside 7 days. Your return policy is the gate on agent checkout.
Generative AI referral traffic to US retail sites rose as much as 4,700% year over year in July 2025. It is a small base, but the slope is what you plan against.
McKinsey puts global agentic commerce revenue at up to 5 trillion by 2030. The range starts at 3 trillion, and the US only figure runs up to 1 trillion.
Three checks that take an afternoon: confirm Agentic Storefronts is on, pull your sessions by referrer report, and get fitment detail into the product text, because agents read text, not your metafields.
Ambaum runs AI readiness audits for Shopify Plus automotive parts merchants. Find out what models see when they look at your store.
