Google Merchant Centre and AI Shopping for AI Mode

Someone is looking for a pair of running shoes.

They might search for "men's running shoes", or ask Google AI Mode which models are best suited to running on roads, offer good cushioning and cost less than £150.

In the second case, the search contains needs, preferences and constraints that the system must interpret to suggest relevant products.

This is one of the major changes brought by AI shopping on Google: users can describe what they are looking for through more detailed requests, compare alternatives and explore specific product features.

AI Mode combines the capabilities of Google's Gemini models with commercial information from Google's Shopping Graph, while new features are bringing these experiences into the Gemini app.

For ecommerce businesses, the quality of product information is becoming increasingly important. Titles, descriptions, attributes and availability all help make a catalogue understandable across different search contexts.

We previously explored the evolution of search in our article on Google AI Mode and its implications for SEO and GEO.

Here, the focus is on how to prepare product data in practical terms.

Why is Google Merchant Centre becoming more important?

Google Merchant Centre collects and organises the information an ecommerce business shares with Google through its product catalogue.

This data powers Shopping experiences and helps make commercial information available across Google's different services.

In its official product data specification, Google explains that accurate, correctly formatted information is also essential for its AI-powered advertising formats and experiences.

Products with incomplete attributes, incorrect identifiers or inconsistent information may encounter eligibility or representation issues.

Google Merchant Centre therefore plays an important role in product discovery, helping users find products that match their needs.

Preparation starts with a catalogue review: approved products, correct GTINs, organised variants, updated prices and information consistent with the website.

These activities already matter for traditional Shopping, and they become even more valuable when users submit more specific requests.

Product feeds: what information helps AI understand what we sell?

A generic title such as "Sports Shoe Model X" communicates very little.

A more detailed description, such as "Men's Running Shoe Model X, High Cushioning, Suitable for Road Running", makes product characteristics and potential uses explicit.

The same principle applies across ecommerce categories.

For clothing, material, fit, colour and size may be important. For electronic devices, compatibility, battery life, dimensions and functionality. For furniture, materials, measurements and intended use.

Google recommends using titles and descriptions that accurately represent each product, keeping the information consistent with the landing page.

Structured attributes such as color, size, material, product_detail and product_highlight also help provide more specific information.

Product feed optimisation should therefore begin with the characteristics that genuinely influence purchasing decisions.

It is important to avoid adding irrelevant keywords or unsupported claims simply to target more searches.

Conversational Attributes: what's new in Merchant Centre in 2026?

In September 2026, Google placed greater emphasis on Conversational Attributes, additional information designed to help AI systems understand more specific aspects of products.

In its official Conversational Attributes guide, Google introduces six optional attributes that can complement existing product feeds.

These include:

• question_and_answer: product-related questions and answers covering compatibility, features and usage.
• document_link: links to relevant documentation, such as manuals and instructions.
• related_product: relationships with compatible products, accessories or items frequently purchased together.
• variant_option: characteristics that distinguish different product variants.

There are also item_group_title, which describes a group of product variants, and popularity_rank, which communicates an item's relative popularity within the catalogue.

Imagine an electronics retailer.

A user might ask AI Mode: "Is this monitor compatible with my MacBook, and does it support a USB-C connection?"

Providing technical specifications, compatibility information and accurate answers can help Google understand the product more effectively.

Google allows merchants to add these attributes through a supplementary data source, the primary feed or the Merchant API.

There is no need to duplicate information already available in description, product_detail or product_highlight.

The objective is to enrich the catalogue with relevant, verifiable data that is genuinely useful to customers.

Images, videos and structured data: what should you optimise?

Shopping search also relies on visual information.

Accurate photography, images showing product variants and content demonstrating how a product is used can help customers evaluate their options.

Google has also introduced the video_link attribute, allowing merchants to associate videos with their products. These assets can be used in free listings and certain advertising formats.

The documentation for video_link recommends videos that demonstrate product features, usage and specific details.

Alongside the feed, the website should communicate consistent information.

Google also recommends using Merchant Centre alongside Product structured data, where relevant, to support a more accurate understanding of commercial information.

Prices, availability, variants and purchasing conditions should remain aligned across the feed, markup and product page.

For AI Mode and AI Overviews, Google also clarifies that no special markup or AI-specific files are required.

Established SEO practices remain relevant, including content accessibility and structured data consistency.

AI performance insights: how can you measure visibility on AI Mode?

Another important development concerns measurement.

On 27 May 2026, Google announced AI performance insights, which subsequently became available in Merchant Centre for accounts in selected markets.

These reports help brands understand how their products appear in conversational shopping searches, particularly across AI Mode and AI Overviews.

According to Google's documentation on AI performance insights, the reports cover several areas.

Share of Voice allows advertisers to compare their brand's visibility with that of competitors included in the report.

Shopping journey insights also distinguish between three stages: Discovery, Evaluation and Ready to buy.

Additional insights cover search terms, shopping intent and the product attributes users are looking for.

For example, if customers frequently search for running shoes offering "maximum cushioning", while a catalogue only includes brand names and model numbers, that may highlight a useful optimisation opportunity.

Availability matters: as of October 2026, these insights are available for English-language queries and Merchant Centre accounts in the United States, Canada, Australia, India and New Zealand. The UK and Italy are not yet included.

The reports also focus on organic traffic across AI experiences, excluding paid advertising traffic.

For UK ecommerce businesses, the priority is therefore to improve product data quality while recognising that these reporting capabilities are not yet available locally.

AI Mode, Gemini and Universal Cart: towards a more connected shopping experience

Developments in Merchant Centre form part of a broader transformation in Google's shopping experience.

On 15 October 2026, Google announced the US launch of Universal Cart, available through Google Search and the Gemini app.

The new basket allows users to save products from different retailers, monitor price changes, receive availability alerts and access simplified checkout options.

Purchasing capabilities also build on the Universal Commerce Protocol (UCP), an infrastructure designed to facilitate commercial interactions between AI agents and merchants.

The rollout remains geographically limited, and access conditions vary depending on the feature.

For brands, the direction is significant: product discovery, evaluation and purchasing can increasingly take place across different Google interfaces.

Maintaining an accurate, updated and structured catalogue therefore provides a useful foundation for future developments in agentic commerce.

Google Merchant Centre and AI shopping: what should you check now?

To prepare an ecommerce business for new Google AI shopping experiences, I would begin by reviewing the existing product catalogue.

The main operational priorities are:

• Feeds and approvals: check Merchant Centre diagnostics, disapproved products and missing attributes.
• Titles and descriptions: include relevant product characteristics, uses and information that supports purchasing decisions.
• Attributes and variants: verify materials, dimensions, compatibility, GTINs and product grouping.
• Conversational Attributes: assess FAQs, documentation and product relationships, starting with the most important categories.
• Visual content: improve images and videos where they can help customers understand a product.
• Data consistency: align prices, stock and product information across feeds, the website and structured data.

For ecommerce businesses operating in supported markets, AI performance insights provide an additional level of analysis to identify terms, intentions and product attributes that may be underrepresented in the catalogue.

Optimisation should follow commercial priorities, beginning with products that generate the most value or offer the greatest opportunities for improvement.

Google Merchant Centre: preparing your catalogue for AI search

New Google Shopping AI experiences are making accurate product descriptions increasingly important.

The Conversational Attributes and AI performance insights introduced in 2026 confirm the direction Google is taking: investing in more detailed information and tools capable of understanding how people search for, compare and choose products.

Visibility across AI Mode and Gemini remains dependent on their respective selection systems, eligibility requirements and feature availability. An optimised feed alone does not guarantee visibility.

For ecommerce businesses, the priority is to build a reliable data foundation that helps Google understand what we sell, which needs our products meet and what information customers need to make a purchase.