Product titles report reveals AI-generated Shopping titles

What changed

Yash Mandlesha shared an observed Product titles report in Google Ads that places customised titles beside the merchant's original wording and performance columns. His post says Google may use either the generated or original title according to its relevance assessment. The preserved interface is labelled beta and describes the report as a sample of titles created by Google AI, so it should not be treated as an exhaustive record of every possible variation.

The image also matters for what it does not show: the table in the supplied capture is empty and says adopted titles appear after sufficient impressions. It verifies the report structure and explanatory labels, not a particular generated title's quality or performance. The columns visible in the image include impressions, product clicks, CTR, cost and average CPC. None of those alone establishes that a wording change improved profit or accurately represented the product.

A retailer could use populated examples to check whether essential product distinctions survive customisation. For instance, a title that blurs the difference between a refill and a complete product could create the wrong expectation even if it attracts clicks. That is a quality-control scenario, not a problem demonstrated in Yash's screenshot. Review the source feed as well: an unclear original title may make it harder to judge whether the generated alternative preserves the intended meaning.

Why it matters for advertisers

Visibility into generated wording can make title automation more reviewable. The useful output is a list of specific issues or promising descriptions tied to identifiable products, rather than a broad conclusion that AI titles are better or worse. Performance comparisons also need context because the versions may be selected for different queries and opportunities.