Google's Deep Research agents now tap private data via API
Google launched Deep Research and Deep Research Max, two autonomous research agents built on Gemini 3.1 Pro that can now search the open web while accessing private enterprise data through a single API call. The agents support Model Context Protocol (MCP) for connecting to third-party data sources and can generate native charts and infographics in research reports. Deep Research prioritizes speed for interactive use cases, while Deep Research Max uses extended test-time compute for exhaustive background analysis. Both are available today in public preview through paid Gemini API tiers.
TL;DR
- →Google released Deep Research and Deep Research Max agents that combine public web search with private enterprise data access via a single API
- →Deep Research optimizes for low-latency interactive use, while Deep Research Max uses extended reasoning for exhaustive background research tasks
- →New agents support Model Context Protocol (MCP) to connect arbitrary third-party data sources and can generate native charts and infographics
- →Available now in public preview through paid Gemini API tiers, built on Gemini 3.1 Pro with reported performance of 93.3% on DeepSearchQA and 54.6% on HLE
Why it matters
This release signals a major shift in how AI agents handle enterprise research workflows. By combining web search with secure access to private data through MCP, Google is positioning autonomous research as a practical tool for finance, life sciences, and market intelligence where accuracy and sourcing are critical. The speed-versus-thoroughness tiering reflects a maturing understanding of how different use cases demand different computational tradeoffs.
Business relevance
For operators and founders, this opens a direct path to embedding research capabilities into products without building custom data connectors. The MCP support means enterprises can plug in existing databases and document systems without API rewrites, while the two-tier model lets teams choose between real-time dashboards and overnight batch analysis depending on use case economics.
Key implications
- →MCP support transforms Deep Research from a web-only tool into a universal data analyst that can query private databases, internal repositories, and specialized services securely
- →The speed-versus-quality tiering reflects a fundamental design choice that other AI agent builders will likely need to address as they scale beyond simple tasks
- →Google is making a direct play for enterprise research workflows in high-stakes industries, positioning its API infrastructure as the backbone for analyst-augmentation products
What to watch
Monitor adoption rates across finance and life sciences to see whether enterprises trust autonomous agents with research tasks where errors carry real costs. Watch for competitive responses from Anthropic, OpenAI, and others on MCP support and multi-source data fusion. Track whether the extended test-time compute approach in Deep Research Max becomes a standard pattern for quality-critical agent tasks.
vff Briefing
Weekly signal. No noise. Built for founders, operators, and AI-curious professionals.
No spam. Unsubscribe any time.
