Gemini 3: Goodbye, organic traffic. Hello, Gen UI.

This week, Google released Gemini 3. If you're not running local models or seeing many sessions referred from the platform, you may think "Oh what a charming dollop of I don't care." Unfortunately, there's a couple of reasons you may need to care.

The self-aggrandizing product announcement starts off simple enough. Blah blah blah big number metrics blah blah blah innovation blah blah blah "And starting today, we’re shipping Gemini at the scale of Google. That includes Gemini 3 in AI Mode in Search with more complex reasoning and new dynamic experiences."

Well, balls.

Gemini 3 in Google Search

Gemini 3 assimilate SERPs starting with AI Mode– and only for Google AI Pro and Ultra subscribers in the U.S. It's also now powering Google Search’s query fan-out technique. The company writes, "Now, not only can it perform even more searches to uncover relevant web content, but because Gemini more intelligently understands your intent it can find new content that it may have previously missed." (They still linked to the query-fanout PDF released in May 2025. If only there was a revolutionary technology to help them write an updated version faster….)

Gemini will also be used to route queries to different models for responses from AI Mode and AI Overviews. Betting money is on Gemini 3 being the transition point to move Web Guides, AI-organized SERPs, from Search Labs to prod. Google promises Gemini 3 will bring generative UI with visual layouts, interactive tools and simulations in AI Mode.

Generative UI

There was actually a flurry of releases across Google for this release. Google Research published a new paper “Generative UI: LLMs are Effective UI Generators– where they described the core principles that enabled implementation of generative UI. Generative UI capabilities will be rolled out as two experiments in the Gemini app: dynamic view and visual layout.

In dynamic view, Gemini designs and codes a fully customized interactive response for each prompt, using Gemini’s agentic coding capabilities. Visual layout appears to be…. Web Guides with flair? Web Guides groups web links in helpful ways — like pages related to specific aspects of your query. Gemini 3's genUI sounds like this– but not limited to web links, instead blending the hodgepodge dynamic view elements.

Generative UI implementation is coming to the Gemini app and Google Search AI Mode. It uses Google’s Gemini 3 Pro model with three important additions:

1. Tool access: A server provides access to several key tools, like image generation and web search. This allows the results to be made accessible to the model to increase quality or sent directly to the user’s browser to improve efficiency.
2. Carefully crafted system instructions: The system is guided by detailed instructions that include the goal, planning, examples and technical specifications, including formatting, tool manuals, and tips for avoiding common errors.
3. Post-processing: The model’s outputs are passed through a set of post-processors to address potential common issues.

"Agentic" shopping

And then there's the shopping. Yes, Gemini 3 is apparently here to make agentic checkout easier for that relative that you don't care too particularly much about. Bright side– it's leveraging the same shopping graph you can include products in from the Merchant Center– "50 billion product listings, 2 billion of which are updated every hour".

But buying from the shopping tab isn't new. Experiments for buying from SERP have been running since at least 2022. The agentic piece seems to be less new technology and more re-purposed Google Duplex– with AI "calling to see if stores nearby have what you're looking for, how much it costs and if there are any special promos". (Pour one out for your friends in retail.) They're also folding in the existing price notifications from Google Shopping– but now with AI!

Feeling the AI fatigue? Don't worry. Sundar Pichai defended the AI bubble by saying if it pops, we're all going down.