Google put names on its next wave of Search ad formats this week, and the more consequential news has nothing to do with the quarter’s revenue line. AI Max, the AI-driven targeting layer Google folded into Search and Performance Max campaigns, is out of beta and now used by more than 500,000 advertisers, the company said. That adoption count matters more than any single earnings figure, because it marks the point where an experimental targeting layer becomes the default way advertisers reach Google’s AI-driven results.

According to a July 23 report from Search Engine Land’s Anu Adegbola, Google SVP and Chief Business Officer Philipp Schindler told investors on Alphabet’s July 22 earnings call that AI Max helps the company match ads to searches too complex or ambiguous for a traditional keyword list. Schindler framed this partly as a revenue-expansion story, saying AI Max opens “billions” of previously unmonetized searches to ads. That is a claim made to shareholders, not an audited figure, and it should be read as one.

Google also told investors that advertisers running AI Max or Performance Max see an average 15 percent lift in conversions or conversion value at a similar return on ad spend, and that Gemini has improved Shopping ad relevance for complex queries by roughly 20 percent. Both numbers are Google’s own, reported on an earnings call with no independent methodology attached. A platform citing its own lift metric while asking advertisers to accept less visibility into ad matching is not a neutral data point.

The more consequential disclosure was Google naming its AI Mode ad formats for the first time. Highlighted Answers is the one worth watching: it inserts sponsored links, tagged as ads, directly into the list-style answers AI Mode generates, rather than placing them in a separate block beside the response. Google said the format is already drawing early user engagement, without saying what that engagement is measured against. Sponsored links now live inside the answer itself, not next to it.

Executives also detailed contextual sitelinks, links drawn from the specifics of a user’s exchange with AI Mode, and Direct Offers, a format that pushes travel and retail promotions to users while they are still weighing options rather than after they have decided. IHG Hotels and Resorts is set to be an early partner for Direct Offers.

The pattern across all three formats is consistent: Google is building ad inventory around intent rather than keyword lists. That trade-off cuts two ways. Product feeds, creative assets and landing pages now carry more of the signal that a keyword list used to carry, which raises the ceiling for accounts that keep those inputs current. It also means advertisers get less insight into which query triggered which ad match, because Gemini performs that matching inside a system Google does not expose in the interface.

Before AI Max starts matching ads to queries an advertiser never typed, three checks are worth running. Confirm product feed fields (titles, categories, availability, price) are complete, since Gemini appears to weight those the way keyword lists were once weighted. Pull the search-terms report weekly instead of monthly, since intent-based matching can drift into adjacent categories faster than manually built campaigns did. Set a brand-safety and negative-keyword review specifically for AI Max line items, given that reaching queries an advertiser did not anticipate is the stated purpose of the format, not a side effect.

For advertisers weighing how far to let AI Max run, the open question is not whether it lifts conversions. Google says it does. The real question is how much visibility into match quality they are willing to give up for that reach, and whether feed quality is strong enough to make that trade worth taking before Highlighted Answers and Direct Offers scale beyond early testing.

Search Engine Land, reported July 23, 2026 by Anu Adegbola, citing Alphabet’s Q2 2026 earnings call held July 22.