Question

How Generative Engines Handle Product Recommendations Versus Informational Queries

LLM answers 4 min read

Ask an AI engine how something works and it explains, citing pages that explain. Ask it which product to buy and it produces a shortlist, and that shortlist changes noticeably from one run to the next. The two kinds of query draw on different evidence and behave differently. Informational answers lean on explanatory pages and cite them. Product answers appear to lean on the wider record of which products are reviewed, discussed and listed, and, at least for Google, on structured product data supplied by merchants. A business that sells products needs to work on both, in different ways.

Two different jobs

An informational query asks the engine to explain: how a heat pump works, what a warranty covers, why a setting matters. The answer is a synthesis, and the useful sources are pages that explain clearly and accurately.

A product recommendation query asks the engine to choose: the best laptop for video editing under a budget, a quiet dishwasher, a reliable running shoe. The answer is a shortlist, and the useful sources are the ones that compare, review and describe specific products.

Where Google shows AI Overviews reflects the difference. Semrush’s AI Overviews study, which tracked more than 10 million keywords in 2025, found Shopping among the industries with the lowest share of keywords showing an AI Overview in November 2025, under 3%. The same study found the share of AI Overviews triggered by commercial and transactional queries growing during 2025, as the informational share fell from 91.3% in January to 57.1% in October. Product queries are a smaller but rising part of where Google shows AI Overviews.

Product recommendations are unstable

A measured feature of product recommendations is how inconsistent they are. A SparkToro study published in January 2026 had 600 volunteers run 12 recommendation prompts 2,961 times across ChatGPT, Claude and Google’s AI. It found less than a 1 in 100 chance that ChatGPT or Google’s AI would return the same list of brands in any two responses if asked 100 times, and closer to a 1 in 1,000 chance of the same order.

Within that variation, established products still recur. Across 994 responses about headphones, brands such as Bose, Sony, Sennheiser and Apple appeared 55% to 77% of the time. For a product business, the realistic goal is a high share of appearances across many answers, not a fixed spot on one list.

Structured product data for Google

Google’s AI optimization guide is specific about product information: “Using products like Merchant Center (such as Merchant Center feeds) and Google Business Profiles can help your products and services to be visible in both AI responses and other Google Search results.” It also mentions Business Agent, “a conversational experience on Google Search that helps customers chat with your brand,” and notes that “protocols like Universal Commerce Protocol (UCP) are emerging that will allow Search agents to do more.”

For informational queries, Google’s guide gives no equivalent feed. Explanatory content is found through crawling and ranking like any other page. For products, a merchant feed is a documented, direct way to supply current product details.

Informational answers cite explanatory pages

For informational queries, AI summaries cite several sources. Pew Research Center found that 88% of the AI summaries in its March 2025 sample cited three or more sources, and that summaries were far more common for question-shaped and longer searches. A product company’s buying guides, how-to articles and technical explanations compete here on the same terms as any publisher’s.

What a product business can do

Query type What helps Where to work
Informational Clear, accurate explanations with evidence and dates Your guides, documentation and help content
Product recommendation Wide, accurate presence in reviews, comparisons and listings; current product data Merchant Center feeds, review programs, independent coverage, comparison content
Both Consistent product names, specifications and claims everywhere Product pages, feeds, retailer listings and press materials

1. Keep product data current and structured. For Google, use Merchant Center feeds, which its AI guide names directly.

  1. Earn independent reviews and coverage. Product shortlists appear to reflect how widely and well a product is discussed. Genuine reviews and editorial coverage build that record; manufactured ones do not, and Google’s AI guide says inauthentic “mentions” do not help.
  2. Publish helpful informational content. Buying guides and explanations that genuinely help a reader decide are citable for informational queries, and they can shape how your category is understood.
  3. Be precise and consistent. Use the same product names, model numbers and specifications on your site, in feeds and on retailer pages, so that every source agrees.
  4. Measure appearance share. Run recommendation prompts many times in each engine and track how often your products appear, rather than reacting to a single answer.

Plan for product recommendations and informational answers as two channels with different inputs. Feed current product data where platforms accept it, earn the independent discussion that product shortlists draw on, and publish explanations good enough to be cited when people are still learning what to buy.