From Chatbot to Agent: Google’s New AI Strategy
Google’s latest I/O keynote made one thing unmistakably clear: Gemini is no longer just a chatbot, it is the backbone of a new agentic era. Instead of waiting for users to type prompts, Google wants Gemini AI agents to quietly live inside every interaction, automating tasks across its ecosystem. The company framed this shift as moving from answers to actions, where AI handles the heavy lifting behind the scenes while users stay focused on goals. This explains why traditional headline features like flashy hardware or big Android overhauls took a back seat. Google’s narrative now revolves around continuity and ubiquity: Gemini embedded in Search, Gmail, Docs, shopping tools, and beyond. The result is an AI-powered environment where you do not start by opening a chatbot; you simply use Google services as usual and let Gemini orchestrate what happens in the background.
Gemini Spark: A Background Agent for Everyday Life
The standout announcement, Gemini Spark, embodies Google’s vision for persistent AI agents. Spark is a cloud-based assistant that runs continuously, connecting to Gmail, Docs, and a growing list of third-party apps like ridesharing, restaurant booking, and real estate platforms. Instead of manually juggling emails, reservations, or itineraries, you delegate multi-step tasks and Spark progresses them while you do something else. It is powered by the lightweight Gemini 3.5 Flash model and built with Google’s Antigravity coding environment, signalling tight integration between consumer tools and developer infrastructure. To address concerns about autonomy, Google introduced Agent Payments Protocol, a system of spending caps, shopping controls, and mandatory user approvals. The company likens Spark to a teenager with a first debit card, suggesting that trust and permissions will loosen gradually. In practice, Spark is Google’s clearest attempt yet at making Gemini an always-on personal operations manager.
Google Search AI Integration: Conversational, Visual, and Sticky
Search remains Google’s crown jewel, and Gemini AI agents are reshaping how people use it. The new intelligent search box allows users to interact more like they would with a conversational assistant, asking follow-up questions, attaching files or videos, and refining queries without starting over. AI Overviews are expanding into full conversational back-and-forth responses, with Gemini generating explanations, visuals, and even short videos directly inside results. This deeper Google Search AI integration keeps users in the search experience longer and reduces the need to click away to multiple sites for basic answers. While that promises faster, more intuitive results for users, it also intensifies concerns among publishers whose traffic may be squeezed by AI summaries. Still, Google’s direction is clear: Search is no longer just a list of links, but a Gemini-powered, context-aware agent that interprets intent, aggregates information, and guides users through complex tasks.
AI-Powered Gmail Features and the New Inbox Workflow
Gmail is becoming a central testbed for Gemini AI agents, turning email from a static archive into a searchable, conversational knowledge base. The new Gmail Live mode lets users speak to their inbox, asking questions like where a flight departs or what deadlines are coming up, and receiving concise answers drawn from messages. Alongside this, AI Inbox upgrades introduce personalized draft replies, one-click task management, and instant access to related Docs and Sheets, effectively blending email triage with workflow automation. These AI-powered Gmail features are designed to reduce manual sorting and replying, letting Gemini handle pattern recognition and summarization. Google emphasizes that user data feeding these responses is not used for training and is traceable via sourcing, so people can see which emails informed a suggestion. Together, these changes recast Gmail as an AI-assisted command center rather than a simple list of messages.
Gemini Ecosystem Expansion: Models, Pricing, and New Interfaces
Underpinning all of this is a broader Gemini ecosystem expansion. Google introduced Gemini 3.5 Flash as the new default model, touting speed, token efficiency, and benchmark gains over earlier Pro versions, with a more powerful Gemini 3.5 Pro on the way. The company is also rolling out a refreshed Gemini app interface called Neural Expressive, bringing richer voice options and animated responses to make interactions feel more personal. On the business side, Google restructured its AI subscription tiers: AI Plus, AI Pro, and AI Ultra, with AI Ultra’s entry point now set at USD 99.99 (approx. RM460) per month, down from USD 250 (approx. RM1150). Instead of counting prompts, usage is now measured by compute, and limits refresh every few hours, with automatic fallback to lighter models. All these changes reinforce Google’s ambition to make Gemini an ever-present agent layer across search, communication, productivity, shopping, and even emerging hardware like audio-focused smart glasses.