Google published one year of AI Mode data this week at I/O 2026, and the headline number is the one nobody planned for. Average AI Mode queries are now triple the length of a classic Search query. The product crossed one billion monthly active users. Queries are more than doubling every quarter. One in six U.S. searches now uses voice or images, with image queries growing more than 40 percent month over month.
A recent post on the Google Blog framed this as natural adoption. From what I see building AI visibility infrastructure at GEOflux.ai, the operational implication is sharper than that. The content most marketing teams produced through 2024 and 2025 was engineered for an interface that is structurally retreating.
The Query Shape Changed Beneath Everything
A three-word keyword and a thirty-word conversational question are not the same data type. They retrieve differently, they get answered differently, and they are scored differently inside the model layer that decides what gets surfaced. Your page, your product description, your blog post was either built to answer the long form, or it was built for keyword density. The two strategies diverge fast at scale.
Planning queries on AI Mode grew 80 percent faster than overall AI Mode queries over the past six months. Brainstorming queries grew 30 percent faster. Users are not asking what a product is. They are asking what to do, where to go, which option to pick, and how to combine three things they have in mind into a coherent next step. The mental model of the user shifted from search to consultation, and the retrieval layer followed.
The cold data we measure across categories at GEOflux makes the gap concrete. Pages that scored well on traditional ranking signals routinely fail to appear in AI-generated answers for the longer-form versions of the same intent. The page exists. The keywords exist. The conversational query lands somewhere else entirely. The retrieval system is asking a different question than the page was built to answer.
What This Forces Teams to Rebuild
Three shifts move from optional to operational this quarter. First, audit your top 50 organic landing pages against the conversational version of the queries they were built to win. Take the keyword the page targets, expand it into a 25-word planning question, run it through AI Mode and ChatGPT, and check whether the page surfaces. Most do not. That is your baseline.
Second, restructure your highest-priority pages around answers to complete decision questions, not around keyword density. The unit of optimization is no longer the keyword. It is the user decision the page helps complete. Include the comparison the user is mentally making. Include the constraint they are working around. Include the next-step they would naturally take. The model is looking for that shape.
Third, accept that multimodal is now structural, not experimental. One in six U.S. searches uses voice or images. Image queries grow 40 percent month over month. If your product cannot be identified from a phone photo, if your service cannot be queried by saying a description out loud, you are excluded from a slice of demand that is compounding by month. Alt text is not enough. Structured product data, image schema, and product images optimized for visual retrieval are now baseline.
The deeper operational shift is that AI Mode and ChatGPT-style interfaces are not a new channel to optimize for in addition to the old one. They are replacing the old query shape. The traffic distribution that built the last fifteen years of SEO consulting is dissolving in slow motion, and the replacement distribution is being decided right now by what the model retrieves on a thirty-word question.
The query type composition is shifting in a way that compounds this gap. Planning queries and brainstorming queries are growing 80 percent and 30 percent faster respectively than overall AI Mode volume. That mix matters because planning and brainstorming sit upstream of the buying decision. The user is not yet ready to transact. They are building the consideration set, weighing options, mapping the decision space. The content that answers planning and brainstorming queries shapes the consideration set for the eventual commercial intent that follows weeks later. Brands absent from the planning queries are also absent from the commercial conversion six steps downstream.
Multimodal compounds again on top of this. Voice search and image search are not separate features. They are different physical interfaces for the same intent. A user who points a phone camera at a product is asking a question the model has to interpret from a pixel grid. A user speaking into AirPods is asking a question the model has to extract from a waveform. Both bypass the classical SEO funnel entirely. The page that wins the text query may not win the voice query or the image query, and the brand strategies that have not been calibrated for all three are leaving compounding share on the table month after month.
Most marketing teams are still budgeting for the three-word keyword universe. The interface moved. The budget did not.
FAQ
How do I optimize content for AI Mode versus classic Google Search?
The shortest answer is to write for the complete decision the user is making, not for the keyword. Take the topic, expand it into a 20 to 30 word planning question that a real user would type, and structure the page to answer that full question with concrete data, comparisons, and a clear next step. Keyword-stuffed pages still rank in classic Search but are routinely skipped in AI Mode retrieval, so the same page often needs both treatments.
Does this mean traditional SEO is dead?
Traditional SEO is not dead, but its share of total search demand is shrinking quickly. Classic Search still drives meaningful traffic, especially for navigational and transactional queries. The fast-growing category is conversational and planning queries, where the retrieval mechanics are different. Keep classic SEO running, but stop assuming it covers your AI visibility.
How do I measure whether my brand is showing up in AI Mode?
Run a regular audit of the conversational versions of your top commercial queries across AI Mode, ChatGPT, and Perplexity. Track how often your brand and pages are cited, what context they appear in, and which competitors get surfaced for the same intent. The mention rate, citation rate, and the kinds of questions you appear for matter more than your classic ranking position at this point.