Google introduced four new AI-powered ad formats at Google Marketing Live 2026, as 92-94% of AI Mode searches now resolve without a click. Here's what each format does and how to prepare.
At Google Marketing Live 2026, Google introduced four new Gemini-powered ad formats: Conversational Discovery ads, Highlighted Answers, AI-powered Shopping ads, and Business Agent for Leads. The context that makes these formats worth understanding rather than ignoring: Google reports that 92 to 94 percent of AI Mode sessions resolve entirely inside Google, with no click to any external website at all.
For a business that has spent years optimizing for clicks, that number should reorganize how you think about paid search. Visibility no longer depends only on ranking near an AI-generated answer. It increasingly depends on appearing inside it.
Conversational Discovery ads answer a person's specific question directly inside an AI Mode response, with Gemini building creative tailored to that exact query rather than serving a pre-built static asset. Someone asking how to make their home smell like a spa might see creative built specifically around that phrasing, generated at the moment of the search.
Highlighted Answers place eligible ads inside AI Mode's list-style recommendation responses, the format Google uses when someone asks for a set of options rather than a single answer. Both formats carry a "Sponsored" label and include an independent, Gemini-written explainer separate from the advertiser's own creative, which Google frames as a transparency and trust measure.
AI-powered Shopping ads go beyond a product image and price. When someone searches for something like an espresso machine, Gemini surfaces the most relevant products and writes a custom explanation of why a specific product might fit that specific search, adjusting the pitch to the query rather than serving the same static listing to everyone.
Business Agent for Leads replaces the static contact form that has long been one of the weakest points in the conversion chain. Instead of filling out a form, a user can open a chat interface directly inside the ad and ask questions, with the AI agent trained on the advertiser's own website content, before a lead is ever handed off.
This is the single most important operational fact in the entire announcement, and it's easy to miss. Highlighted Answers eligibility is not purely a bidding auction. Relevance and content quality determine whether your business appears, which means the lever that matters most going into these formats is the quality and specificity of your existing content and structured product data, not simply how much you're willing to spend.
Review your product feed and ad copy for genuine specificity. Vague descriptions, phrases like "high-quality product for every need," give Gemini very little to build a tailored, query-specific ad from. Specific, concrete product and service descriptions are what these formats generate creative from, which means the content you already have on hand is doing more work than it used to.
If you run lead generation rather than ecommerce, review whether your website content can actually support a Business Agent conversation. The agent answers using your site as its information source, so thin or outdated website content becomes a direct liability the moment a prospective customer starts asking it real questions.
Separate your reporting for AI Mode traffic from standard Search traffic where possible. Because AI Mode sessions may create different user journeys than a traditional search click, qualified lead rate, chat quality, and post-click behavior deserve their own scrutiny rather than being judged against the same benchmarks as conventional Search conversions.
Google's Search and advertising revenue reached $60.4 billion in the first quarter of 2026, up 19 percent year over year, running directly counter to predictions that AI-generated answers would hollow out Search's advertising business. Google isn't losing the ad business to AI. It's rebuilding the ad business around AI, and these formats are the infrastructure for that shift.
Not a completely different strategy, but a different point of emphasis. Structured product data, specific rather than vague copy, and website content substantial enough to support an AI agent's answers all matter more than they did under a purely bid-based system.
Availability varies by format. Business Agent for Leads is in open beta for US advertisers on English-language accounts. Conversational Discovery ads and Highlighted Answers are in testing on US mobile and desktop, with AI-powered Shopping ads expected to expand later this year.
Not necessarily. Because these formats create different user journeys, particularly Business Agent for Leads replacing a static form with a conversation, qualified lead rate and post-click behavior should be evaluated on their own terms rather than measured against historical Search benchmarks.
This connects directly to the broader shift we've covered in what actually determines AI citation eligibility and why local businesses are losing customers to AI recommendations without realizing it. If you're not sure whether your product feed and website content are ready for how paid search is changing, that's exactly what our paid media team audits as part of a free Growth Gap Analysis.
How We Work
