Google Merchant Center this month introduced a beta capability that allows advertisers to populate product listings by running an AI-powered scan of their website, bypassing the need to prepare a structured data feed before launching Shopping campaigns. The feature, labeled "Use AI to add products - beta" inside the "Add product source" section of the platform, was first spotted and reported by Saurabh on X and confirmed by Search Engine Roundtable on May 4, 2026.

According to Search Engine Roundtable, Google's description of the feature reads: "Use Google's AI to automatically scan your website and add your products. It's the easiest way to get started and helps reduce errors from manual entry." The option sits alongside two existing methods - uploading a product file manually and connecting a Google Sheets spreadsheet - but carries a notable technical distinction: it performs only a one-time scan rather than continuously synchronising with the merchant's website.

What the interface shows

The "Add product source" screen, visible in screenshots circulated on X and LinkedIn, presents three side-by-side options. The AI option occupies the leftmost card and displays a "BETA" tag. Beneath the description, it is marked as a "One-time scan." The middle card, "Add products from a file," and the right card, "Use Google Sheets," both carry an "Automatically updates" label with a sync icon - a distinction that carries practical weight for advertisers comparing how fresh their product data will remain after setup.

The screenshots show the configuration targeting India as the market and English as the language, with products set to appear across "all marketing methods." A "Continue" button completes the flow. The screenshot was shared by Google Ads consultant Adriaan Dekker on LinkedIn on May 5, 2026, who noted the rollout was confirmed as of that date.

How the AI scan works

According to Adriaan Dekker's LinkedIn post, the capability lets advertisers "create product listings using AI" with "no product feed required," and the mechanism is a one-time site scan. This places the feature in the same broad category as Merchant Center's existing automatic import functionality, which was introduced as part of Merchant Center Next and also reads product data from a merchant's website. However, the two differ in a critical way: automatic import claims to refresh every 24 hours, while the new AI beta option makes no such promise and instead positions itself explicitly as a one-time operation.

That one-time nature has direct implications for product feed management. Standard feeds built through file upload or Google Sheets update automatically whenever the source data changes - prices, availability, and descriptions can propagate into Merchant Center without manual intervention. An AI-scanned catalog, by contrast, captures a snapshot of the website at a point in time. Merchants who update pricing frequently or who operate catalogs where stock levels change day to day would be working from data that degrades in accuracy as soon as the scan concludes.

The target use case, according to Dekker's post, is "quick setup or smaller catalogs." Advertisers without a structured feed setup, or those running websites where generating a programmatic feed is technically difficult, could use the AI scan to get products live faster than the file-based route allows.

Context in Merchant Center's data source history

The introduction of this beta option arrives against a backdrop of sustained changes to how Google Merchant Center handles product data ingestion. Google completed its global rollout of Merchant Center Next in August 2024, with automatic website feed population as one of its headline simplification features for new merchants. The platform's emphasis at that time was on reducing the technical burden of feed creation, and automatic import was the primary mechanism for doing so.

Since then, however, independent analysis documented in January 2026 found that Merchant Center's automatic import was failing to meet its own stated 24-hour update frequency. Emmanuel Flossie of FeedArmy documented a nine-day gap between the platform's declared refresh cycle and actual product data timestamps in at least one account - an inconsistency that has direct consequences for Shopping ads that depend on accurate inventory and pricing signals.

That gap matters more acutely as Google's retail ad surfaces multiply. Product feed data now powers free listings, AI Mode, the Gemini shopping experience, virtual try-on in Google Lens, and Business Agent, in addition to the traditional Shopping ads surfaces. A feed that is nine days stale does not just affect a single campaign - it feeds into multiple surfaces simultaneously.

Against that context, a feature designed as a one-time scan occupies an unusual position. It addresses the barrier of initial setup but does not engage with the ongoing data freshness problem that has drawn criticism. Merchants who adopt it for quick launch would need to return and update their product source - or switch to one of the automatically updating methods - once their catalog evolves.

How this fits into feed method options

Google's Merchant API, announced as generally available in August 2025, establishes the programmatic route for managing Merchant Center product data at scale, with the Content API for Shopping scheduled for shutdown on August 18, 2026. For large retailers managing thousands of SKUs and automated pricing systems, the API path is the appropriate one. File uploads and Google Sheets serve mid-tier merchants comfortable managing a structured export from their e-commerce platform. The new AI beta sits at the other end of that spectrum - oriented toward the smallest, least technically equipped segment.

Product feed method selection has historically carried performance consequences beyond setup convenience. A January 2026 analysis covered by PPC Land found that merchants using automatic imports surrender control over product identifier assignment to Google's detection algorithms, which can produce ID mismatches between Merchant Center and the website's remarketing tags. Dynamic remarketing campaigns, in particular, rely on consistent product IDs between the feed and the site's event tracking. When those IDs differ, display campaigns cannot match user behavior to specific products, producing generic ads or mismatched product tiles that waste ad spend.

The AI beta scan, by operating as a one-time import rather than a continuous sync, would likely produce a similar ID assignment dynamic. Google's AI reads the site and generates product identifiers; the merchant does not control what those identifiers are. For a small advertiser running basic Shopping ads without dynamic remarketing, the practical consequences of this would likely be minimal. For any merchant planning to layer in audience-based retargeting or dynamic product ads at a later stage, the ID consistency question would need to be revisited.

The broader AI automation direction

The timing of this beta places it within a sequence of AI-driven product and advertising features Google has been deploying across its retail stack throughout 2025 and 2026. On April 30, 2026, Google extended AI Max to standard Shopping campaigns, introducing text customization that draws on Merchant Center feed attributes to generate ad copy tailored to individual search queries - including product characteristics like fabric softness, material durability, or fit for apparel items. That feature reads the feed; the quality of its outputs depends on how much structured attribute data the feed contains.

An AI-scanned catalog may not produce the same attribute depth as a manually prepared feed with dedicated fields for material, size, color, and product category. The AI would detect what is visible and parseable on the website, but product detail pages vary enormously in how they expose structured data. A feed built from a scan of a page where price and title are prominently marked up may differ substantially from a feed built from a page where the relevant attributes are embedded in less structured HTML.

The Shopping tab's free listing grid began mixing paid and organic results from April 20, 2026, with sponsored tiles appearing inside what had historically been an organic-only display environment. Shopping ads have also been available inside AI Mode since February 2026, following an announcement on February 14, 2026, where the format was described as appearing labeled as sponsored within conversational search sessions. The surface area that product data must now serve is considerably wider than it was when Merchant Center was primarily a tool for managing Shopping ad feed submissions.

Merchant Center product onboarding options compared

Three methods are currently available under the "Add product source" flow as visible in the interface screenshots. The AI beta performs a one-time scan and does not update. The file upload method requires technical knowledge to create a properly formatted product file, but it automatically updates when the source file is refreshed. The Google Sheets method uses a spreadsheet containing product titles, descriptions, prices, and additional attributes; it also updates automatically. Only the AI option carries the "BETA" label at this time.

Google has repeatedly tightened Merchant Center product data requirements over the past two years. A March 2026 policy update mandated that "out of stock" products must display a visibly greyed-out, non-clickable buy button on the landing page rather than a hidden one - a requirement that places fresh emphasis on consistency between feed data and live page content. In January 2026, Google announced that merchants must use separate product IDs when the same item differs between online and in-store versions, effective March 2026, affecting all Local Inventory Ads users. These compliance requirements apply regardless of how product data entered Merchant Center - whether through a file, a spreadsheet, or an AI scan.

Significance for the marketing and advertising community

PPC Land has tracked Merchant Center data quality and reliability issues extensively throughout 2025 and 2026, covering problems ranging from the nine-day automatic import lag to smart cropping distorting product images without a user-accessible opt-out, and the "Product Page Unavailable" error that halts Shopping ads even when the merchant's site is fully operational. The new AI beta does not address any of these underlying data quality concerns.

What it does address is the initial friction of getting products into Merchant Center for advertisers who have no existing feed infrastructure. For small businesses operating custom-built websites, or advertisers launching a new account quickly, the AI scan option removes a genuine barrier. But the one-time nature of the scan means it is not a substitute for a properly maintained feed - it is a starting point that would need to be replaced or supplemented as the catalog grows and data freshness becomes relevant.

The beta label signals that Google is still evaluating the feature. No rollout timeline or general availability date has been announced.

Timeline

Summary

Who: Google, through its Merchant Center platform; first documented by Saurabh on X and reported by Search Engine Roundtable (Barry Schwartz) on May 4, 2026; further confirmed by Google Ads consultant Adriaan Dekker on LinkedIn on May 5, 2026.

What: A beta feature labeled "Use AI to add products - beta" now appears in the Google Merchant Center "Add product source" interface. The feature performs a one-time AI-powered scan of a merchant's website to detect and import products, including titles, descriptions, prices, and other attributes, without requiring the advertiser to prepare or upload a product feed file.

When: The feature was first publicly spotted on May 4, 2026 and confirmed as rolling out as of May 5, 2026. No general availability date has been announced.

Where: Inside Google Merchant Center, under the "Add product source" section, where advertisers select how to populate their product catalog. The screenshots show the interface configured for India as market and English as language.

Why: The feature addresses the technical barrier of initial product catalog setup for advertisers who lack structured feed infrastructure. According to Dekker's post, it is best suited for quick setup or smaller catalogs. It sits within a broader Google strategy of reducing onboarding friction in Merchant Center while simultaneously expanding the number of surfaces that rely on Merchant Center product data - including Shopping ads, free listings, AI Mode, the Gemini shopping experience, and connected television formats.

Share this article
The link has been copied!