Google Search AI overhaul launches with Gemini 3.5 Flash and new

Google announced that Search’s AI Mode will now run on the Gemini 3.5 Flash model and that users can create “information agents” to track web changes, marking the biggest UI shift since the search box debuted over 25 years ago【1】.

AI‑first Search box and Gemini 3.5 Flash

Google’s revamped search box now expands automatically to accommodate longer, conversational queries, removing the need for users to toggle an AI mode before typing【2】. The box also offers AI‑generated query suggestions that go beyond traditional autocomplete, helping users craft more nuanced searches【2】. Under the hood, the AI Mode is powered by Gemini 3.5 Flash, the latest iteration of Google’s Gemini family announced at I/O 2026【1】. This model replaces the earlier Gemini 3.5 Flash‑based “Gemini Omni” rollout for subscribers and powers new features like Gemini Spark, a 24/7 personal AI agent integrated with Workspace tools【1】.

Information agents and broader AI integration

Google introduced “information agents” that run continuously in the background, scanning the web for updates related to a user’s predefined queries and delivering synthesized alerts—an evolution of the 2003 Google Alerts service【2】. The agents are initially focused on “information” tasks but are positioned to expand into more specialized mini‑apps. Google also expanded Personal Intelligence in AI Mode to nearly 200 countries and territories across 98 languages, extending the reach of its AI‑driven search experience【1】.

Competitive context

The shift from static “ten blue links” to AI‑driven interactive results mirrors broader industry moves toward conversational search, but Google’s integration of persistent agents and a unified Gemini model across Search, Workspace, and the Gemini app creates a tighter ecosystem than rivals that rely on separate large‑language‑model offerings. By unifying the model stack, Google can leverage the same underlying reasoning and multimodal capabilities—such as the ability to generate cohesive videos from mixed media inputs in Gemini Omni—across its product line, potentially accelerating feature parity with competitors that currently stitch together disparate AI services【1】.

What to watch

• Global rollout timeline – Google will continue expanding the new search box and agents to all markets where AI Mode is already available, starting with India【1】.
• Gemini Spark integration – Monitoring how the 24/7 personal AI agent is adopted within Gmail, Docs, and Slides will indicate the depth of AI penetration in everyday productivity tools.
• Competitor responses – Watch for announcements from Microsoft and other search providers on whether they will introduce comparable persistent agents or unified model architectures.

The overhaul signals Google’s intent to make AI the default lens through which users access the web, turning search from a passive list of links into an active, personalized information hub. Whether the new agents will become indispensable or remain a niche feature remains to be seen.