Recently, tech giant Google officially launched two of its own research agents, Deep Research and Deep Research Max, based on its powerful Gemini 3.1 Pro model. These tools aim to replace humans in intensive information collection and analysis tasks through automated workflows, marking a shift of AI assistants from simple Q&A interactions to complex long-term task execution.
Currently, these two agents are available for preview through the paid version of the Gemini API to developers worldwide. They can efficiently retrieve public web information and also access private enterprise databases through newly supported protocols, providing users with professional in-depth reports with complete source references.
Two versions for flexible adaptation, balancing speed and research depth
To meet different office needs, Google has introduced two versions of the agent. The standard version, Deep Research, focuses on low latency and high response speed, making it ideal for integration into interactive user interfaces that require immediate feedback, helping users quickly sort out research leads.
In contrast, the Deep Research Max version pursues maximum comprehensiveness and quality. It can use extended computational power during testing to perform multiple iterations of reasoning, searching, and refining, acting like a seasoned analyst who works overnight in the background to complete complex due diligence and generate highly valuable final summaries.
Supports MCP protocol, native charts aid data visualization
A major highlight of this update is the introduction of the Model Context Protocol (MCP). This means companies can directly connect the agent to specific financial, market, or industry-specific databases, breaking the previous limitation that AI could not access private "data islands," achieving seamless integration between public network information and private knowledge.
In addition, the system has strong native visualization capabilities, capable of generating high-quality charts and infographics directly from the research results. Before the research begins, users can also collaborate with AI to create detailed search plans, ensuring precise research direction. These combined features indicate that the barriers to professional intelligence analysis will be further reduced.
