Google reimagines Search with AI agents and generative interfaces

The Search redesign is arguably as consequential as the model launches. The classic ten-blue-links surface is being supplanted by an AI Mode that handles multi-step tasks, runs agents on the user's behalf, and synthesizes generative UIs tailored to each query — booking interfaces for travel queries, comparison dashboards for product research, summary cards for news. AI Overviews already reaches 2.5 billion users, and the new agent surfaces are being layered on top of that distribution.

The mechanics are tied to Gemini 3.5 Flash powering AI Mode, with the Antigravity Managed Agents runtime handling task execution behind the curtain. Google's framing is that Search becomes the place users complete tasks, not just look things up — a direct shot at ChatGPT, Perplexity, and Claude's nascent web experiences.

The publishing industry's reaction is unambiguously alarmed. TechCrunch's headline declared "Google Search as you know it is over," reflecting concern that generative interfaces capture more queries inside Google's surface without clickthroughs. SEO operators are scrambling to understand how content gets surfaced (or doesn't) when the interface is generated on the fly rather than ranked.

The competitive context: OpenAI's ChatGPT search and Perplexity's Comet have validated the AI-answer-engine pattern, but neither has Google's 90%+ search market share to leverage. By marrying agents to the dominant query surface, Google is doing what only Google can — using existing distribution to outflank entrants who have superior products but no users. Watch for further publisher litigation, antitrust angles, and a probable wave of "Gemini-optimized" content guidance from Google in the next 30 days.