The Blinking Cursor That Changed the World Just Changed Itself
For 25 years, the Google search box was the most consequential rectangle in the history of computing. A long, slender white bar, centred on a white page, with a blinking cursor and two buttons beneath it. Billions of people learned to compress their information needs into two or three keywords and trust the algorithm to do the rest. "World Cup." "Best pizza near me." "How to tie a tie." The box did not ask for context. It did not ask follow-up questions. It returned links and expected you to do the rest of the thinking. On May 19, 2026, at Google I/O in Mountain View, California, Google announced that bar is gone. In its place is what Google's vice president and head of Search, Liz Reid, called the Intelligent Search Box — the biggest upgrade to the company's iconic search interface since its debut over 25 years ago. The change is not cosmetic. It is architectural. And its implications extend far beyond how you find a pizza recommendation.
The numbers that drove Google to act are significant. AI Mode, which launched in the United States at I/O 2025, has surpassed one billion monthly users in its first year, with AI Mode queries having doubled every quarter since launch. AI Overviews, the lighter-weight AI summaries, now reach more than 2.5 billion monthly users, and overall search query volume hit an all-time high last quarter. Sundar Pichai, Google's CEO, used these figures as evidence that AI features are additive rather than cannibalistic to search usage, telling the I/O audience: "When people use our AI-powered features in search, they use search more." The data gave Google the confidence to make a change that a company with five billion daily search users might otherwise be too cautious to attempt. When your incumbent product is growing faster with AI than without it, redesigning the interface around AI stops being a gamble and becomes the obvious move.
What the Intelligent Search Box Actually Does
Google redesigned the search box to give searchers more space to ask longer, deeper queries. The search box will continue to expand as the user enters the query or prompt. There is an AI-powered suggestion that Google's head of Search, Liz Reid, said "goes beyond autocomplete." Users can now search with text, images, files, videos, or their Chrome tabs — putting Google's most powerful AI tools right at the user's fingertips and making it easier to ask complex questions. The capability to search using an open Chrome tab is the feature that will matter most to professional users. Instead of copying text from a document, switching to a browser tab, and pasting a search query, a user can simply point the search box at the tab they are already reading and ask a question about it. The friction reduction is small per interaction and enormous at scale — across billions of searches per day, removing one extra step from the process of finding information represents a structural improvement in how efficiently human attention gets converted into useful answers.
Google is also merging its AI Overviews and AI Mode features into a single, seamless search flow, eliminating the friction that previously forced users to choose between a traditional results page and an AI-forward experience. The merger matters because it removes the two-speed search interface that has created confusion since AI Overviews launched. Previously, users encountered a hybrid experience: some queries returned AI summaries, others returned traditional links, and the logic governing which mode applied was opaque to users. The unified interface means every search begins with the Intelligent Search Box and proceeds through a single AI-native experience that can escalate to deeper AI Mode when the query warrants it. The user no longer has to decide which mode to use. The system decides based on what the query needs.
From Keywords to Conversations: Why This Shift Has Been 25 Years in the Making
For 25 years, Google's iconic search box was a long, slender bar where people typed in keywords like "World Cup." Over the past three years, artificial intelligence allowed people to type in longer, more complex questions like "Who are the top 24 teams in the World Cup, and what chance does the United States have of advancing?" The AI shift inspired Google to overhaul its search bar for the first time since 2001, with the box getting bigger and more interactive so people can ask even longer questions and upload photographs and videos into queries. People can now also ask follow-up questions with a chatbot on Google's main search page. This is the linguistic shift that underpins everything else. Keyword search was a compression protocol — users learned to strip their questions down to the fewest possible words that would return useful results. Conversational search is the opposite: it rewards specificity, context, and the kind of natural language that people use when talking to a knowledgeable person. The search box is expanding physically because the queries it is now designed to receive are longer. It is expanding functionally because the capabilities it offers — image input, file upload, video query, follow-up questions — require interface surface that a two-button rectangle cannot provide.
The company expects capital expenditures of approximately $180 to $190 billion in 2026 — roughly six times the $31 billion it spent four years ago — largely to support the infrastructure required for this AI transformation. When asked about the future of traditional search, Sundar Pichai was direct: "Search is the most used AI product in the world." That reframing is deliberate and consequential. By describing Google Search as an AI product — rather than a search engine that has added AI features — Pichai is repositioning Google's core business in the context of a competitive landscape that has spent three years arguing that ChatGPT, Perplexity, and Claude represent existential threats to keyword search. If search is already the most used AI product in the world, the competitive narrative shifts: Google is not a legacy search engine trying to catch up to AI. It is the AI interface that already serves five billion users daily, and it is now building the interface that makes that advantage permanent.
The Search Agents: When Google Searches For You
Google will also offer digital assistants, known as agents, to automate searches so that someone who may be apartment hunting can be notified of a new listing without opening a real estate site like Zillow. The search features are powered by Gemini 3.5 Flash, a new AI model that Google said had improved on creating software code and performing autonomous tasks, worked faster and was less expensive to run than comparable models. The apartment hunting agent is a deliberately approachable example of a capability with much broader implications. An agent that monitors the web for conditions matching a user's criteria and delivers results without requiring an active search represents the inversion of the original search model. Traditional search is pull: the user initiates a query, the engine returns results, the user decides what to do. Agentic search is push: the user defines a goal, the agent monitors for relevant developments, and delivers results when conditions are met — whether or not the user is currently thinking about the topic.
The commercial implications for every industry that currently depends on Google Search as a discovery mechanism are profound. Real estate portals, travel booking sites, job boards, e-commerce marketplaces — every vertical that currently captures user attention by being ranked highly in search results faces a future in which Google's agents potentially intermediate that discovery relationship. If a user delegates their apartment search to a Google agent, they may never visit Zillow or Rightmove — the agent will surface the relevant listings directly within the search interface. The advertising model, the affiliate model, and the SEO industry that has grown up around Google's original keyword search architecture were all built on the assumption that users would click through to third-party sites. The Intelligent Search Box and its associated agents are being designed for a world in which that assumption is progressively less reliable.
Generative UI: Search That Builds Experiences, Not Just Returns Links
Google announced what it calls "generative UI" — the ability for search to dynamically build custom widgets, interactive visualisations, and even mini applications in real time, tailored to a user's specific question. Reid offered a concrete example: a user could ask "How do black holes affect spacetime?" and receive an interactive visual in an AI Overview that brings the concept to life — not a link to an explainer article, but a purpose-built interactive experience generated specifically for that query. Generative UI is the feature that will be hardest for competing AI search products to replicate because it requires not just a language model capable of generating accurate answers but a rendering infrastructure capable of turning those answers into interactive experiences at query speed and at web scale. Google's advantage here is not just Gemini — it is the combination of Gemini with the infrastructure that has served trillions of search queries and the engineering culture that has spent 25 years optimising every millisecond of search response time.
The competitive moat that Google is building with the Intelligent Search Box is therefore not primarily a model quality moat. It is an infrastructure moat — the combination of Gemini's capability, Search's distribution, the agent architecture's monitoring capability, and generative UI's rendering infrastructure represents a full-stack AI search product that no competitor currently assembles in a single offering. OpenAI's SearchGPT has the language model and some web access. Perplexity has the conversational interface and source citations. Neither has 25 years of search infrastructure, five billion daily users, or the advertising ecosystem that makes Google's search economics self-funding. The 25-year-old search box that Google has just replaced was the interface of the keyword era. Its replacement is the interface of the AI era — and Google is betting $190 billion of annual capital expenditure that it has built it first and built it right.