Google Merchant Center AI Product Import Beta: No Feed Required
Google Merchant Center added an AI product import beta on May 5, 2026, allowing the system to scan a website and automatically generate product listings without a structured XML or CSV data feed. For merchants who previously needed feed management tools or developer resources to get products into Shopping campaigns, this removes the primary onboarding barrier. The feature is in beta; validation against existing feed data quality is required before treating AI-generated listings as the primary source.
Google Merchant Center AI Product Import Beta: No Feed Required
Google Merchant Center added an AI product import beta feature on May 5, 2026, that scans a website and automatically generates product listings without requiring a structured data feed. Previously, adding products to Google Shopping required creating and maintaining an XML or CSV product feed -- a technical step that required either developer resources, a feed management platform, or a CMS plugin. The AI import scans the product pages directly and extracts the product data needed for Shopping listings. For merchants who previously avoided Shopping campaigns due to feed complexity, this removes the primary barrier to entry.
Key takeaways
- Google Merchant Center AI product import (beta, May 5, 2026) scans website product pages to generate Shopping listings without requiring a structured XML/CSV feed -- reducing setup from days to hours for eligible merchants.
- The feature is in beta and availability is limited -- not all accounts have access. Check Merchant Center under Data Sources for the AI import option.
- Validate AI-generated listings against your actual product catalog before treating them as the primary source: verify product title accuracy, pricing, availability, and image selection across a sample of SKUs.
- For accounts already using manual feeds, AI import is an alternative onboarding path, not a replacement for verified feed data -- run both in parallel initially to measure listing coverage and quality.
- For campaign performance monitoring after enabling AI import, connecting Google Ads to Prooflytics surfaces conversion rate changes by product group so AI-generated versus manual-feed listing performance can be compared.
What the AI product import beta does
Google Merchant Center product feed: a structured file (XML or CSV) containing product data in Google's required format -- title, description, price, availability, GTIN, image URLs, and other required attributes. Traditionally, this feed is generated by an ecommerce platform, maintained by a developer, or managed by a third-party feed management service.
The AI product import replaces feed creation with automated website scanning. Google's system crawls the merchant's product pages, extracts the product data it finds (title, price, images, availability indicators), and generates Merchant Center product listings from that data. The advertiser reviews and activates the generated listings rather than creating or uploading a feed file.
This mirrors how dynamic remarketing and Dynamic Search Ads already work -- both use website scanning to identify content -- but applies the same approach to the product data onboarding workflow in Merchant Center specifically.
Who benefits most from AI product import
Merchants without feed infrastructure. Small and mid-sized merchants who sell on their own website but have never run Shopping campaigns because feed setup required technical resources they don't have. AI import removes the technical prerequisite.
Merchants testing new channels. Brands with an existing feed for one marketplace (Shopify feed for Google, or Amazon feed) who want to test other Shopping surfaces without building additional feed infrastructure. AI import provides a fast path to testing without feed development work.
Merchants with feed latency issues. Accounts where the current feed pipeline has a 24-48 hour delay between product updates on the website and Merchant Center listing updates. AI import could provide more real-time product data if the crawl frequency is faster than the feed update cycle.
Accounts already running Shopping campaigns with well-maintained, validated feeds should evaluate AI import as a parallel test, not an immediate replacement. Feed data that has been validated against product catalog accuracy is the more reliable starting point for active campaigns.
Stop stitching platform exports together
Every channel in one brief — plus the memory of what each one actually drove.
14 days free · no credit card
How to access and set up AI product import
Check for beta access
AI product import is in beta and access is not universal. In Google Merchant Center, navigate to Data Sources (previously called Feeds). If your account has beta access, you will see an option to add a data source via AI scanning in addition to the standard feed upload options. If the AI import option is not visible, the account does not yet have access.
Configure the AI scan
When setting up AI product import:
Website URL: provide the root URL of the website containing the product catalog. The AI scanner will crawl from this starting point.
Crawl scope: specify whether the scanner should crawl the entire site or specific subdirectories. For large catalogs, limiting the initial scan to a specific product category or URL pattern reduces the initial review workload.
Review period: after the initial scan, Google generates a preview of the product listings it extracted. Review this preview before activation -- check for product titles that are truncated or garbled, prices that were not correctly extracted, availability status that doesn't match the actual inventory, and images that are incorrect (Google selecting secondary images rather than the main product image).
Validate listing quality against your catalog
Before activating AI-generated listings for active campaigns, validate a sample of the generated listings against your actual product catalog:
- Select a random sample of 20-30 product listings across different categories
- Check product title: does it match the product name as you sell it? Is it truncated at an appropriate length?
- Check price: does it match the current selling price, including any active discounts?
- Check availability: is the availability status (in stock / out of stock / preorder) correct?
- Check primary image: is the primary product image selected (not a banner, lifestyle photo without the product, or thumbnail)?
- Check GTIN/MPN: if your products have GTINs, are they extracted correctly? Missing GTINs affect Shopping eligibility for certain product categories.
Issues found in validation should be noted and compared against the existing feed (if one exists) to assess whether AI import meets the same quality bar.
Run AI import in parallel with existing feed initially
For accounts already using a manual or platform-generated feed, add AI import as a secondary data source rather than replacing the primary feed immediately. Merchant Center allows multiple data sources for the same account. Running both in parallel allows:
- Comparison of listing coverage: does AI import capture the same SKU count as the manual feed?
- Quality comparison: do AI-generated titles and descriptions match the feed data quality standards?
- Performance comparison: do campaigns using AI-generated listings achieve similar CTR and conversion rates as campaigns using manual feed listings?
After validation, if AI import quality is equivalent or better, consider migrating to AI import as the primary source to reduce feed maintenance overhead.
Attribution models tell you which channels got credit; incrementality testing tells you which channels actually drove decisions — geo holdout testing is the most practical way to find out, as described in Geo Holdout Testing: How to Measure True Marketing Incrementality.
What to watch for in early testing
Image selection errors. AI scanning may select secondary images, promotional banners, or context images rather than clean product images on white or neutral backgrounds. Product images are a primary driver of Shopping ad CTR and the primary determinant of whether a product image meets Google's Shopping image policy requirements.
Price extraction accuracy. Sites with dynamic pricing (cart-based discounts, logged-in pricing, regional pricing) may show inconsistent prices to the scanner versus to customers. AI import reflects the price shown to the unauthenticated scanner, which may not match the price shown to customers after login or region detection.
Title quality for SEO-optimized product pages. Product pages optimized for organic search sometimes use title formats that are more descriptive than they are useful for Shopping listings. AI import will extract the page title or H1, which may be better for organic SEO than for paid Shopping performance. Standard Shopping best practice is product titles that include product type + brand + key attributes (color, size, material).
Coverage gaps. AI scanning may miss products on pages with non-standard markup, products loaded via JavaScript after page load (if the scanner uses a basic crawler without JavaScript rendering), or products in paginated catalog pages beyond the first several pages.
Bottom line
- Google Merchant Center AI product import (beta, May 5, 2026) scans product pages and generates Shopping listings without requiring a structured feed -- reducing setup time from days to hours for merchants who previously couldn't or didn't run Shopping campaigns.
- The feature is in beta with limited availability -- check Data Sources in Merchant Center to see if your account has access.
- Validate a sample of AI-generated listings against the actual catalog before activating for campaigns: check titles, prices, availability, and primary image selection.
- For accounts with existing feeds, run AI import in parallel initially to compare coverage and quality before considering migration.
- For Shopping campaign performance visibility after enabling AI import, see how Prooflytics connects Google Ads to surface product-level conversion data in a daily briefing.
- See independent reviews of Google Shopping and ecommerce analytics tools on G2.
Frequently asked questions
Does Google Merchant Center AI import replace product feeds?+
Not definitively -- it is an alternative path for product data ingestion, particularly for merchants who previously could not create or maintain feeds. For accounts with well-functioning, validated feeds, AI import is a parallel testing option. For new accounts or merchants who have avoided Shopping due to feed complexity, AI import removes the main setup barrier. Whether it replaces feeds depends on whether AI-generated listing quality meets the standards required for the specific account and campaign type.
Is the AI product import feature available for all Merchant Center accounts?+
No. The feature launched in beta on May 5, 2026, and access is limited. Check your Merchant Center Data Sources section for the AI import option. Google has not published a specific timeline for broader availability. Accounts that don't see the option should monitor Google Merchant Center updates for when the feature exits beta.
How accurate is product data extracted by AI import?+
Accuracy varies by site structure. Well-structured product pages with clear H1 titles, visible prices, availability indicators, and structured data markup (Schema.org Product) tend to produce higher-quality AI import results. Sites without structured data markup or with complex pricing logic may require more manual correction after the initial AI scan. Validation against a sample of listings before campaign activation is required regardless of site structure.
Can AI import handle large product catalogs?+
The beta feature handles catalogs of varying sizes, but very large catalogs (tens of thousands of SKUs) may require extended processing time and produce higher volumes of listings to review. For large catalogs, a phased approach -- testing AI import on a specific product category first before expanding to the full catalog -- is recommended to manage the validation workload.
What happens if the AI import picks up incorrect product data?+
Incorrect listings generated by AI import can be edited manually in Merchant Center or by updating the source product pages with clearer structured data markup. For persistent quality issues, structured data markup (Schema.org Product schema on product pages) helps the AI scanner identify the correct data fields. Adding Schema.org markup for price, availability, name, image, and brand is the highest-priority remediation step for accounts with AI import data quality issues.
Stop stitching platform exports together
Every channel in one brief — plus the memory of what each one actually drove.
14 days free · no credit card
Continue reading
Google AI Max for Shopping Campaigns: Text Customization, URL Expansion, and Format Selection
Google rolled out AI Max to Shopping campaigns on April 30, 2026, introducing three new capabilities: dynamic text customization (ads adapt copy to conversational queries), final URL expansion (the system suggests additional landing pages), and format selection (advertiser control over ad layout variants). For Shopping advertisers, this requires reviewing feed data quality and URL structure before activation, as AI Max uses the product feed as the primary creative input for generated ad variants.
Google Product Packs: How to Win E-Commerce Traffic in 2026
Google Product Packs have become the primary e-commerce traffic channel for merchants who appear in them -- and a dead zone for those who don't. Analysis of 63,000+ merchants shows specific visibility gaps between winners and losers. The differences are not in bidding. They are in feed quality, structured data, and pricing competitiveness.
Google Ads API Now Reports Product-Level Metrics for PMax, Video, and Demand Gen
Google Ads API adds product-level cost and conversion data for Performance Max, Video, Demand Gen, and App campaigns on June 15, 2026. Agencies and in-house teams can now build automated product-ROI alerts without manual UI extraction.
Google AI Max for Brand Campaigns: What the Early Performance Data Shows
AI Max expands Google search targeting beyond keyword lists using landing pages and site content as signals. Early testing shows 35% lower ROAS than traditional match types in one study and $100 per conversion versus $44 for phrase match in another. Here is what to check before enabling AI Max on your brand traffic.