Google AI Mode Search Features: Home-Page Buttons and Ask Advisor

Three confirmed moves in the past week point the same direction: Google is actively steering users into Google AI Mode search features and turning its own reporting tools into agentic conversations.

A home-page button test acknowledged by a Google VP, the launch of the Ask Advisor agent across Ads and Analytics, and a parallel ad build-out inside ChatGPT are not three unrelated stories. They are one trend read, and for developers tracking where search traffic goes, the signal is that the answer layer is becoming the default surface rather than a feature bolted onto a list of links.

What are Google AI Mode search features?

Before the trend read, the plain definition. AI Mode is Google’s conversational, Gemini-powered search experience: instead of returning a page of ten blue links, it synthesizes an answer and lets you follow up in natural language, the same way you would with a chatbot. The “features” Google is now surfacing on the home page are shortcuts into specific AI Mode tasks. Four are worth knowing:

• Create images — generate a picture from a text description straight from Search, without opening a separate image tool.
• Ask about files — upload a document (a PDF, a spreadsheet) and ask questions about its contents, getting a synthesized answer grounded in what you uploaded.
• Brainstorm — an open-ended, back-and-forth mode for exploring ideas, planning, and comparisons where the goal is a conversation rather than a single lookup.
• In-field AI Mode — an AI Mode entry point built into the search box itself, so a typed query can resolve as a generated answer rather than a results page.

In short, these are task-shaped doorways into generative search. The rest of this piece is about where Google is putting those doorways and what that placement signals.

Home-page buttons push Google AI Mode search features front and center

Google confirmed it is testing new buttons directly beneath the search box on its home page, and the buttons open AI Mode rather than a standard results page. In a limited desktop test for signed-out users, the familiar Google Search button is replaced by Create images, Ask about files, and Brainstorm, with an AI Mode entry point inside the search field itself (Search Engine Land, August 2026).

Robby Stein, Google’s VP of Product, called it “a small test on desktop to help people find new things they can do with Search,” adding there is “no impact to how the core search box works” (Search Engine Land, August 2026). Read plainly, Google is spending its most valuable pixels, the home page, to route first-time and unauthenticated users toward generative search before they type a keyword.

And the surface being promoted is not niche. AI Mode already crossed 1 billion monthly users, and AI Overviews reached 2.5 billion monthly users as of Google I/O 2026 (Dataconomy, June 2026).

Ask Advisor turns reporting into an agentic conversation

One agent across four tools

On the measurement side, Google launched Ask Advisor, a unified Gemini-powered agent spanning Google Ads, Analytics, Merchant Center, and Display & Video 360, announced at Google Marketing Live 2026 (Search Engine Land, August 2026).

The pitch is agentic, not just conversational: marketers can diagnose a performance drop, set up campaigns, generate creative, and build custom visuals with natural-language commands without switching tools (Google Ads & Commerce Blog, 2026).

Why it matters for anyone who lives in dashboards

Two pieces are worth flagging. Google Analytics is adding AI Overviews on its homepage that summarize what changed since your last login and recommend next steps, and Ads and Analytics are gaining AI dashboards built from raw data with text prompts (Search Engine Land, August 2026).

The same interaction model Google is pushing on searchers, ask a question and get a synthesized answer, is now the interface for your own analytics. Investigating a conversion drop that once meant checking Ads, then Analytics, then feed health separately collapses into one query.

ChatGPT’s ad build-out is the same signal from OpenAI

The trend is not Google-only. In the same window, ChatGPT Ads rolled out conversion-optimized oCPC campaigns in beta, made Automatic Advanced Matching the default for new web pixels, and began testing single-brand product carousels at the bottom of answers for advertisers including L.L.Bean, VistaPrint, and Walmart (Search Engine Land, August 2026). Automatic Advanced Matching also switches on for existing pixels on August 17, 2026 unless advertisers opt out (Search Engine Land, August 2026).

Taken with Google’s moves, the pattern is clear: the answer engine is being wired for commerce and measurement at the same pace it is being promoted to users. Both companies are building the monetization and reporting scaffolding that a search product needs once the answer, not the link list, is the destination.

What this signals for developers tracking search traffic

The through-line is that the generative surface is graduating from an experiment to the front door, and every layer around it, home-page routing, ad formats, and reporting, is being rebuilt to match. For teams shipping client sites, three positions follow.

• Expect more traffic to resolve inside the answer, not on your page. Promoting AI Mode from the home page accelerates the zero-click pattern already visible in the data. We covered why a #1 ranking can still go unseen in AI search visibility versus organic rankings, and nothing here reverses that; it deepens it.
• Treat citation, not placement, as the durable asset. The commerce layers growing inside both engines are a separate pipeline from the one that selects sources, a split we broke down in the ChatGPT ads study. Paying to appear in an answer does not make you the answer.
• Instrument for the agentic surface. When your own analytics answers in prose and recommends next steps, the questions you ask it become the interface. Ask precise questions, and verify the synthesized numbers against the raw reports before you act on them.

None of this changes the technical work that earns visibility: machine-readable structure, specific answer-first prose, and a credible place in the citation graph. It changes where that work pays off. If you want that baseline encoded once and applied to every page rather than left to each writer, the rest of our engineering write-ups go deeper on the build patterns behind it.