Does AI Max for Shopping replace standard Shopping campaigns?

No. AI Max layers onto standard Shopping campaigns as a one-click upgrade. It adds reach into conversational queries, AI Mode, and AI Overviews without removing your existing product targeting or bidding controls.

Which Merchant Center attributes matter most for AI Max?

Material, fit, pattern, size_type, age_group, gender, color, item_group_id, product_highlight, and well-populated product_detail fields. AI Max reads context beyond the title to match long-tail conversational queries.

Do I need to change my titles for AI Max to work?

Not necessarily. AI Max text customization can rewrite titles to match shopper intent, but it still relies on the underlying feed data. Rich attributes give the model more accurate raw material than a stuffed title ever could.

Can I turn off Final URL Expansion in AI Max for Shopping?

Yes. Google documents that you can disable Final URL Expansion at any time, which restricts ad delivery to Shopping ads only and keeps shoppers on the product URL declared in your feed.

AI Max for Shopping: Feed Attributes That Drive AI Mode Reach

AI Max for Shopping changes how Google reads your product feed. For years, Merchant Center fields outside title, description, and price were treated as filters. With AI Max, those same fields decide whether your products surface for conversational queries inside AI Mode and AI Overviews.

That makes feed depth a reach lever, not a hygiene chore. This guide walks through which attributes matter, why, and how to prepare an existing feed for the upgrade.

What changed with AI Max for Shopping

Google launched AI Max for Shopping campaigns as a beta in April 2026, then expanded it at Google Marketing Live in May 2026. The Google Ads blog frames it as a one-click upgrade that adds three capabilities to standard Shopping campaigns:

Text customization that generates ad copy aligned to shopper intent
Final URL Expansion that matches shoppers to the most relevant landing page
Optimal Format Selection that picks between text and Shopping ad formats

Underneath those three features sits a more important shift. Google explicitly states the system “leverages your Merchant Center feed including details like fabric softness, material durability and fit to understand the context of their products.” That sentence is the entire feed strategy in one line.

Why feed depth now drives conversational reach

Standard Shopping campaigns matched ads to queries by mapping shopper words to feed attributes. AI Max removes that constraint. The system can now respond to natural language queries that no feed attribute would match directly.

Example. A shopper types “comfortable loose-fitting linen trousers for summer.”

Under standard matching, the campaign needed those exact words somewhere in the feed. Under AI Max, the system reads:

material: linen
fit: relaxed
season or product_highlight: summer or warm weather
product_detail entries for breathability or weight

It assembles that context, decides your product is a match, and writes ad copy explaining why. The query never had to match a single attribute literally.

The attributes that matter most for AI Max

Google has not published a ranked list, but the documentation and customer examples point to a clear hierarchy.

Tier one: soft product attributes

These are the fields that historically went unfilled because they were optional and did not affect approval.

material
pattern
fit
size_type
age_group
gender
color

AI Max draws on these to interpret descriptive queries. A feed with all seven populated has dramatically more conversational coverage than a feed with only title, price, and image.

Tier two: structured detail fields

product_highlight
product_detail
additional_image_link

Product highlights act as bullet evidence the model can paraphrase. Product detail is structured key-value data, which AI Max can quote directly inside conversational ad formats.

Tier three: grouping and identity

item_group_id
gtin
mpn
brand

These do not change conversational matching directly, but they let AI Max pick the right variant for a query and prevent cannibalization across near-identical products.

How to audit an existing feed for AI Max readiness

Run the audit in this order. Each step takes minutes per product and unlocks measurable reach.

Step one: count completion rates

Pull a feed snapshot and calculate the percentage of products with non-null material, pattern, fit, and product_highlight. Most feeds sit below 40 percent on these fields. Aim for 90 percent before flipping the upgrade.

Step two: fix marketing fluff in descriptions

AI Max parses descriptions for factual context. “Premium quality fabric for the modern professional” gives the model nothing. “100 percent combed cotton, 220 GSM, machine washable, slim fit” gives the model five facts it can reuse.

Step three: validate item_group_id consistency

Every variant in a group should share an item_group_id. Missing or inconsistent groupings make AI Max guess which variant to show.

Step four: confirm landing pages match Final URL Expansion intent

AI Max can substitute a more relevant URL when Final URL Expansion is on. Walk your top 20 products and check whether category pages, variant pages, and bundle pages all return clean, fast, in-stock results.

Step five: re-run the feed audit

A clean feed amplifies AI Max. A feed with disapprovals, GTIN mismatches, or image issues amplifies the wrong things.

AI Max for Shopping performance benchmarks

Google has published three reference numbers from the AI Max beta. They are worth reading literally, with the feed context behind each result.

Advertiser	Result	Source
Lufthansa Group	24 percent increase in return on ad spend	Google blog, GML 2026
IKEA	65 percent lift in non-branded clicks, 28 percent incremental ROAS boost	Google blog, GML 2026
Across advertisers	15 percent more conversions at similar ROAS when adopting AI Max or PMax	Google blog, GML 2026

Both Lufthansa and IKEA operate catalogs with deep attribute structures already in place. The lesson is not that AI Max produces those lifts automatically. The lesson is that AI Max produces those lifts when the underlying feed has enough structured context to work with. A thin feed under AI Max will look like a thin feed under standard Shopping, with more wasted impressions.

Where Final URL Expansion changes the work

Final URL Expansion is the AI Max feature most likely to surprise merchants. When enabled, AI Max can route a shopper away from the link declared in your feed if a more relevant URL exists on your site.

That is useful when a shopper query matches a category page better than a single product. It is risky when the alternative pages are slow, low-converting, or out of stock. Two safeguards apply.

Audit category and bundle pages for inventory and load time.
Use the AI Max control to disable Final URL Expansion if you cannot guarantee site-wide quality.

How AI Max changes feed operations going forward

The feed is no longer a static delivery file. It is the raw material AI Max uses to write ad copy, choose formats, and decide whether you appear in conversational answers at all.

That means feed management work shifts in three ways.

Soft attributes become priority-one fields, not optional polish.
Description quality affects ad copy directly, not just landing-page relevance.
Feed freshness affects how quickly AI Max can adapt copy to seasonal or stock changes.

Teams that already run disciplined feed operations will see the largest lift. Teams that treat the feed as a one-time export will lose ground to competitors who do not.

Where to go next

If you are catching up on every announcement from Google Marketing Live 2026, start with the Google Marketing Live 2026 Shopping recap. If your feed needs structural cleanup before you enable AI Max, work through the Google merchant feed field audit for new products. If you sell on Shopify, see how AI Max reads your Shopify product catalog for platform-specific guidance.

AI Max for Shopping makes feed depth a campaign lever. Merchant teams that treat the feed as a structured data product, not a passive export, will see the upside first.