The Story

Google has launched Deep Research and Deep Research Max, two new AI agents designed to automate complex research workflows. As per Google's official announcement, the agents use Gemini 3.1 Pro to generate detailed research reports from a combination of web sources, uploaded documents, and enterprise data systems.

The Details

• Unified research engine: Both agents run on Gemini 3.1 Pro and are integrated into Google's research stack, replacing earlier preview versions.
• Multi-source input: Users can combine open web search, uploaded files, and Model Context Protocol (MCP) servers to generate comprehensive reports.
• Flexible data control: Users can choose to include or exclude web data, enabling private, enterprise-only research workflows when needed.
• Improved performance: Google reports gains in retrieval and reasoning benchmarks compared to earlier versions and competing models.
• Enterprise integrations: Google is working with partners like PitchBook, S&P, and FactSet to integrate proprietary financial data into the research pipeline.

Why it matters

Google is targeting one of the most valuable AI use cases: automating high-effort knowledge work. By turning research into a structured, API-accessible workflow, the company is pushing AI deeper into enterprise decision-making. This could significantly reshape industries like finance, consulting, and legal services, where time-intensive research has long been a core function.