Deep Research Workflow

Starting from Gatsbi 3.0, Gatsbi Writer includes a built-in Deep Research Workflow that allows users to conduct web-based academic research by simply providing instructions.

This marks a major shift from tool-assisted writing to a fully autonomous research agent.

How It Works

Step 1: Provide Instructions

You start by entering your research instructions, for example:

“Analyze how UK SME support policies influence firm innovation performance.”

Instructions can be broad, exploratory, or incomplete.

Step 2: Automatic Research Question Formulation

Gatsbi interprets your instructions and automatically:

• Clarifies the research scope
• Refines implicit research questions
• Identifies key research dimensions

This ensures the task is academically meaningful and researchable.

Step 3: Methodology Proposal

Based on the research context, Gatsbi proposes suitable methodologies, such as:

• Policy and institutional analysis
• Qualitative or conceptual synthesis
• Comparative or framework-based analysis

This step is especially optimized for social sciences and humanities research.

Step 4: Web-Based Deep Research

Gatsbi then conducts web-search-driven deep research, including:

• Searching academic literature and policy sources
• Collecting relevant evidence from the web
• Filtering and organizing materials by relevance
• Aligning evidence with research questions

This replaces manual literature searching and early-stage review work.

Step 5: Evidence Synthesis and Conclusion Derivation

Finally, Gatsbi:

• Synthesizes collected evidence into presentation-ready forms (visuals and narratives).
• Constructs coherent academic arguments
• Identifies patterns, tensions, and implications
• Derives structured research conclusions

The result is a research-ready narrative rather than a simple summary.

Why This Matters

Before 3.0, users needed to bring their own research materials.

From Gatsbi 3.0 onward, you can:

• Start from instructions only
• Let Gatsbi conduct web-based deep research
• Focus on interpretation, refinement, and originality

This significantly lowers the barrier for high-quality academic research, especially in fields where evidence is dispersed and qualitative.