Deep Research transforms Gemini into a personal research assistant

Google recently unveiled Deep Research, a new Gemini Advanced feature that combines Google's search expertise with Gemini's reasoning capabilities and 1M token context window to deliver comprehensive research reports on practically any subject. Available for English-speaking users in the desktop and mobile web experiences as "Gemini 1.5 Pro with Deep Research", this feature processes research questions and, as a first step, designs a "multistep research plan" that users can revise and modify as they see fit.

Once the research plan is approved, Gemini collects and analyzes relevant information gathered from the Internet, constantly iterating and refining the gathered insights in a process designed to resemble what a human would do when researching a topic: Gemini performs an initial search, finds relevant information, and then performs additional searches based on what it learns. When done, Gemini presents its findings as an in-depth report that can be exported to a Google Doc and includes links to the original sources. Any follow-up questions or additional modifications to the report can be done right in the chat window.

According to Google, since Deep Research is all about saving time and performing research that would take hours in minutes, the feature is ideal for anyone, from graduate students preparing a presentation on a specific subject to entrepreneurs who want to speed up the process of doing a comprehensive competitor analysis and marketers who need to accelerate research into recent campaigns for future benchmarking.