Research Agent
The Research Agent block conducts comprehensive information gathering and analysis by systematically searching, evaluating, and synthesizing data from multiple sources.
Overview
Research Agent specializes in deep information gathering, combining web search, database queries, and document analysis to build comprehensive knowledge bases and answer complex questions that require multiple sources and analytical thinking.
Research Planning: Break down research questions into searchable components and information needs
Source Discovery: Search across web, databases, documents, and APIs for relevant information
Content Evaluation: Assess source credibility, relevance, and information quality
Knowledge Synthesis: Combine findings into coherent insights and comprehensive reports
How It Works
graph TB
A[Research Query] --> B[Query Decomposition]
B --> C[Multi-Source Search]
C --> D[Web Search]
C --> E[Database Query]
C --> F[Document Analysis]
D --> G[Source Evaluation]
E --> G
F --> G
G --> H[Information Synthesis]
H --> I[Research Report]Configuration
Information Sources
Available data sources for research:
- Web Search: Google, Bing, DuckDuckGo for current information
- Academic Databases: PubMed, ArXiv, Google Scholar for scholarly sources
- News Sources: Reuters, AP, industry publications for current events
- Internal Databases: Company knowledge bases, CRM data, documentation
Search Strategies
Methods for finding relevant information:
- Broad to Narrow: Start with general searches, then drill down into specifics
- Multi-Angle: Approach the topic from different perspectives and disciplines
- Temporal: Search across different time periods for historical context
- Cross-Reference: Verify information across multiple independent sources
Quality Criteria
Standards for evaluating source reliability:
- Authority: Source expertise and credentials
- Accuracy: Fact-checking and verification against known information
- Currency: Information recency and relevance to current context
- Objectivity: Bias assessment and balanced perspective
Output Format
How research findings are presented:
- Executive Summary: Key findings and recommendations
- Detailed Report: Comprehensive analysis with supporting evidence
- Source Bibliography: Complete reference list with credibility ratings
- Structured Data: Organized facts and figures for further processing
Use Cases
- Market Analysis: Comprehensive industry research including trends, competitors, and opportunities
- Academic Research: Literature reviews and evidence gathering for scholarly work
- Due Diligence: Background research on companies, individuals, or investment opportunities
Example Workflow
[Research Question] → [Research Agent] → [Knowledge Graph] → [Expert Summary] → [Presentation]A market research request: "Analyze the competitive landscape for AI-powered customer service tools"
Research Process:
- Query Decomposition: AI customer service, market size, key players, trends, pricing
- Source Discovery: Industry reports (Gartner, Forrester), company websites, news articles
- Information Gathering: 50+ sources across 3 categories (vendors, analysts, users)
- Source Evaluation: Rate credibility, check publication dates, verify claims
- Synthesis: Market map, competitive matrix, trend analysis, recommendations
Output: Comprehensive 15-page report with market size ($2.8B), top 10 competitors, growth trends (23% CAGR), and strategic recommendations
Best Practice: Always verify important facts across multiple independent sources. Maintain clear source attribution and credibility ratings to help users assess information quality.
When to Use This vs Other Blocks
| Block | When to Use |
|---|---|
| Research Agent | Comprehensive multi-source research requiring analysis and synthesis |
| Agent | Simple information lookup or single-step question answering |
| Knowledge Graph | Organizing and connecting research findings after collection |