How to get your products into Google AI Overviews.

The slow way: logged-out queries and a screenshot folder

The manual baseline is worth one afternoon. Open Google in a logged-out private window and run ten buyer-style queries for your category — "best [category] for [use case]," feature comparisons, a competitor query, a price-band query. For each, record three things: did an AI Overview appear at all (coverage varies heavily by query type), is your PDP among the cited sources, and which domains won the citations. Screenshot the overview, because the same query can return a different summary next week and you will want the receipt.

Then comes the part the spreadsheet does badly. AI Overviews and the shopping carousel under it fail for different reasons — the written answer draws on your PDP and its schema, the carousel draws on Merchant Center — so a "we're not showing up" note tells you nothing about which fix to ship. Diagnosis means cross-checking indexation in Search Console, validating JSON-LD per page, and auditing the feed per SKU. Most teams do it once, fix one thing, and never re-test on a cadence that would show whether the fix worked.

The eCommerce Insights way

1. Baseline in 45 seconds, then continuously. The free AI Overviews citation checker reports the controllable layer for any PDP — crawler admittance, schema completeness, answer coverage, entity clarity, plus a web-grounded Gemini probe. The paid tracker then runs your query set against all three Google AI surfaces on every refresh, per product.
2. Clear the indexation gate first. Not indexed means not citable, full stop. The indexation cross-check catches noindex tags, crawl blocks, and server errors before they read as mysterious citation losses — see fix robots.txt blocking AI crawlers for the bot-by-bot version.
3. Complete the Product JSON-LD. Missing gtin, brand, or offers fields make a SKU ambiguous, and ambiguous SKUs get skipped for complete competitors. Schema fixes ship as reviewable diffs — the workflow in fix product schema for AI search.
4. Write the answer block. AI Overviews quotes and paraphrases pages that answer the buyer's question in sentences, not only in a spec table. Rewrite the PDP opening to answer the query directly, add FAQ pairs with FAQPage schema, and keep claims quotable — the technique is covered in optimize content for AI search.
5. Fix the feed for the carousel, then watch the deltas. Carousel absence is usually a Merchant Center problem — missing GTIN, disapproval, thin attributes — not a PDP problem, and the dashboard says which. Week-over-week deltas across all three surfaces separate a real citation win from the feature's normal churn.

What "good" looks like

Directional reads for a mid-market D2C catalog, from eCommerce Insights tracking as of mid-2026 (illustrative):

Two cautions. First, AI Overview coverage is query-dependent — a citation rate only means something against the subset of your queries that trigger the feature at all. Second, the written answer and the carousel are scored separately for a reason: a product can win one and lose the other, and the fixes do not overlap.

Related jobs