Felo AI Intelligent Search Agent Creation Manual: Complete Guide to Building Professional AI Assistants

Ken Fong July 16, 2025

Introduction

Felo AI Agent is a powerful tool that combines user-defined instructions with specific intelligent tools, designed to help users quickly and accurately obtain the information they need. By integrating artificial intelligence technology with the actual needs of businesses or individuals, it can efficiently and stably perform tasks in various scenarios, greatly enhancing work efficiency and the convenience of information retrieval.

This comprehensive manual will guide you through the complete process of creating a professional AI agent using Felo AI's platform, from basic setup to advanced features implementation. Specifically, this manual provides a comprehensive 5-step guide to creating Felo AI Intelligent Search Agents - custom AI assistants that leverage specialized knowledge sources for professional and business applications. Learn to build agents that integrate custom instructions with intelligent tools to deliver precise, knowledge-based results.


Quick Overview:

  • Step 1: Configure basic agent information
  • Step 2: Define custom prompts with knowledge integration
  • Step 3: Select AI model (Claude 4 Sonnet)
  • Step 4: Choose appropriate tools for intelligent search
  • Step 5: Save and deploy your agent

Quick Start Checklist

Before you begin:

  • Access to Felo AI platform 
  • Clear understanding of your agent's purpose
  • Knowledge domain identified
  • Target user audience defined
  • Desired output format determined

Introduction

Felo AI offers two distinct methods for creating AI agents to enhance your workflow and research capabilities. This comprehensive manual specifically focuses on Intelligent Search Agent creation - one of the two available methods on the Felo AI platform.

Understanding Felo AI's Agent Creation Methods

Felo AI provides users with two powerful approaches to agent creation:

  1. 🔍 Intelligent Search Agent (Beta) - Uses custom instructions and specific knowledge sources to search for results
  2. 🔄 Multi-step Search Agent - Automatically plans multiple search steps based on the query to obtain results

This manual covers the Intelligent Search Agent creation process, which allows for full customization with user-defined prompts, roles, and tool selection.

What is a Felo AI Intelligent Search Agent?

A Felo AI Intelligent Search Agent is a sophisticated tool that combines user-defined instructions with specific intelligent tools and knowledge sources, designed to help users quickly and accurately obtain targeted information. By integrating artificial intelligence technology with custom instructions and specialized knowledge bases, it can efficiently perform tasks in various scenarios, greatly enhancing work efficiency and the convenience of information retrieval.

Felo Agent Store Integration

All agents created through both methods (Intelligent Search Agents and Multi-step Search Agents) are accessible through the Felo Agent Store (https://felo.ai/agents), which serves as a centralized hub for both pre-built templates and user-created agents.

Prerequisites

Before starting, ensure you have:

  • Access to Felo AI platform
  • Basic understanding of AI prompt writing
  • Clear vision of your intended agent's purpose and functionality
  • Understanding that you'll be creating an Intelligent Search Agent (Beta feature)

🔧 STEP 1: Configure Basic Agent Information

1.1 Agent Name (Required)

  • Purpose: Provide a concise and clear identifier for your Intelligent Search Agent
  • Guidelines:
    • Keep it descriptive and professional
    • Maximum 70 characters
    • Follow consistent naming pattern
  • Standardized Examples:
    • "Market Research Intelligence Agent"
    • "Academic Research Intelligence Agent"
    • "Business Strategy Intelligence Agent"
  • Action: Enter the name in the "Agent Name" input field

1.2 Agent Description

  • Purpose: Briefly describe the main functions and problem-solving capabilities of your Intelligent Search Agent
  • Guidelines:
    • Maximum 500 characters
    • Focus on core functionality and knowledge sources
    • Use consistent professional format
  • Standardized Examples:
    • "Utilizes specialized knowledge sources to conduct comprehensive market analysis and generate strategic business insights"
    • "Leverages academic databases and research sources to provide structured literature reviews and research summaries"
    • "Integrates business intelligence sources to deliver strategic recommendations and competitive analysis"
  • Action: Fill in the "Agent Description" text area

1.3 Input Box Placeholder

  • Purpose: Guide users on how to interact with your Intelligent Search Agent
  • Guidelines:
    • Maximum 500 characters
    • Provide clear instruction on expected input format
    • Use consistent style across all agents
  • Standardized Examples:
    • "Enter your market research topic and specific industry focus"
    • "Describe your research question and academic domain"
    • "Specify your business challenge and strategic objectives"
  • Action: Complete the "Input Box Placeholder" field

📝 STEP 2: Define Core Instructions - Custom Prompt (Critical Step)

2.1 Understanding Custom Prompts for Intelligent Search Agents

The custom prompt is the most critical component for Intelligent Search Agents as it determines how your agent will interpret queries, utilize knowledge sources, and deliver results. A well-crafted prompt for an Intelligent Search Agent includes four key elements:

2.2 Intelligent Search Agent Prompt Structure Framework

A. Role (Persona)

Define the agent's expertise and knowledge domain

Standardized Examples:

  • "You are a senior market intelligence analyst with access to specialized business databases and industry knowledge sources"
  • "You are an academic research specialist with expertise in scientific literature analysis and access to comprehensive research databases"
  • "You are a strategic business consultant with deep industry knowledge and access to competitive intelligence sources"

B. Task & Knowledge Integration

Clearly specify what the agent should accomplish using its knowledge sources

Standardized Examples:

  • "Analyze market data from specified knowledge sources and generate comprehensive intelligence reports with strategic insights"
  • "Search academic databases systematically and synthesize findings into structured research summaries with proper citations"
  • "Utilize business intelligence sources to provide strategic recommendations and competitive analysis for decision-making"

C. Context/Constraints & Knowledge Sources

Provide background information, limitations, and specify knowledge source utilization

Standardized Examples:

  • "Focus on recent data from the last 2 years, prioritize credible business sources, and maintain analytical objectivity"
  • "Emphasize peer-reviewed sources, ensure academic rigor, and provide comprehensive literature coverage"
  • "Leverage industry-specific knowledge bases, focus on actionable insights, and maintain strategic perspective"

D. Output Format & Intelligence Delivery

Specify desired response structure for intelligent search results

Standardized Examples:

  • "Present findings in structured business intelligence format with executive summary, detailed analysis, and strategic recommendations"
  • "Provide academic summary with literature review, key findings, research gaps, and suggested future directions"
  • "Structure responses with strategic overview, competitive analysis, market opportunities, and implementation recommendations"

2.3 Intelligent Search Agent Writing Guidelines

  • Maximum 10,000 characters
  • Emphasize knowledge source utilization
  • Include instructions for handling multiple search steps
  • Specify how to synthesize information from various sources
  • Consider edge cases in knowledge retrieval

2.4 Standardized Custom Prompt Examples

Example 1: Market Research Intelligence Agent

You are a senior market intelligence analyst with access to specialized business databases and industry knowledge sources. Your expertise lies in synthesizing market information from multiple knowledge sources to provide comprehensive business insights.

Your task is to utilize your knowledge sources systematically to analyze market data and generate comprehensive intelligence reports with strategic insights based on the user's specified industry or business challenge.

When processing queries, you should:
1. Search through relevant business knowledge sources and databases
2. Cross-reference information from multiple specialized sources
3. Analyze trends and patterns from your knowledge base
4. Synthesize findings into actionable business intelligence

Your response must include:
1. Executive summary of key market intelligence
2. Detailed analysis with supporting data from knowledge sources
3. Competitive landscape overview (minimum 3 key players)
4. Market trends and strategic opportunities identification
5. Strategic recommendations based on intelligence gathered
6. Clear citations of knowledge sources utilized

Focus on recent data from the last 2 years, prioritize credible business sources, and maintain analytical objectivity. Present findings in structured business intelligence format with executive summary, detailed analysis, and strategic recommendations.

Example 2: Academic Research Intelligence Agent

You are an academic research specialist with expertise in scientific literature analysis and access to comprehensive research databases. Your expertise lies in synthesizing academic information from multiple knowledge sources to provide structured research summaries.

Your task is to utilize your knowledge sources systematically to search academic databases and generate comprehensive literature reviews with proper citations based on the user's specified research question and academic domain.

When processing queries, you should:
1. Search through relevant academic knowledge sources and databases
2. Cross-reference information from multiple scholarly sources
3. Analyze research trends and patterns from your knowledge base
4. Synthesize findings into structured academic summaries

Your response must include:
1. Executive summary of key research findings
2. Detailed literature review with supporting evidence from knowledge sources
3. Research methodology overview (minimum 3 key approaches)
4. Current research trends and knowledge gaps identification
5. Academic recommendations based on literature gathered
6. Clear citations of knowledge sources utilized

Emphasize peer-reviewed sources, ensure academic rigor, and provide comprehensive literature coverage. Provide academic summary with literature review, key findings, research gaps, and suggested future directions.

Example 3: Business Strategy Intelligence Agent

You are a strategic business consultant with deep industry knowledge and access to competitive intelligence sources. Your expertise lies in synthesizing strategic information from multiple knowledge sources to provide comprehensive business recommendations.

Your task is to utilize your knowledge sources systematically to analyze business intelligence and generate strategic recommendations with competitive analysis based on the user's specified business challenge and strategic objectives.

When processing queries, you should:
1. Search through relevant business intelligence sources and databases
2. Cross-reference information from multiple strategic sources
3. Analyze competitive trends and patterns from your knowledge base
4. Synthesize findings into actionable strategic intelligence

Your response must include:
1. Executive summary of key strategic intelligence
2. Detailed competitive analysis with supporting data from knowledge sources
3. Strategic landscape overview (minimum 3 key competitors)
4. Business opportunities and strategic recommendations identification
5. Implementation recommendations based on intelligence gathered
6. Clear citations of knowledge sources utilized

Leverage industry-specific knowledge bases, focus on actionable insights, and maintain strategic perspective. Structure responses with strategic overview, competitive analysis, market opportunities, and implementation recommendations.

🤖 STEP 3: Select AI Model

3.1 Model Selection for Intelligent Search Agents

  • Current Default: Reasoning Model Pro (Claude 4 Sonnet) - upgraded from Claude 3.7 in late 2024
  • Beta Limitation: The reasoning model for this beta version cannot be changed
  • Status: Toggle is greyed out and non-selectable during beta phase
  • Future Availability: "Basic version will be open soon" - indicating more model options coming
  • Model Evolution: Felo AI has enhanced its AI capabilities by upgrading from Claude 3.7 to Claude 4 Sonnet, providing superior reasoning and language understanding
  • Optimization: Specifically optimized for knowledge-based search and analysis tasks
  • Enhanced Features:
    • Improved coding and reasoning capabilities
    • More refined language understanding
    • Better creative capabilities for content generation
    • Enhanced precision in responding to complex queries
  • Action: The "Reasoning Model Pro" is automatically enabled and cannot be modified
  • Recommendation: The fixed Pro model provides optimal performance for Intelligent Search Agents

3.2 Why Reasoning Model Pro for Intelligent Search Agents

  • Knowledge Integration: Superior ability to synthesize information from multiple sources
  • Analytical Precision: Enhanced reasoning for complex business and academic analysis
  • Multilingual Support: Better cross-language capabilities for global knowledge sources
  • Professional Output: Improved formatting and structure for business-ready results
  • Beta Optimization: Model is specifically configured for beta testing and optimal performance

3.3 Future Model Options

  • Basic Version: Will be available soon for users who need lighter processing
  • Model Flexibility: Future updates will allow model selection based on specific needs
  • Performance Tiers: Different models may offer varying capabilities and processing speeds

🛠️ STEP 4: Tool Management and Selection for Intelligent Search Agents

4.1 Accessing Tool Management

  • Action: Click the "Manage Tools" button below the Tools bar
  • Interface: A modal window will appear showing all available tools with toggle switches
  • Principle: Select tools that enhance knowledge source utilization and intelligent search capabilities


4.2 Complete Available Tools for Intelligent Search Agents

🔍 Search Internet

  • Function: Access real-time information to complement knowledge sources
  • Intelligent Search Use: Verify and update knowledge base information
  • Best Practice: Use for current data validation and trend verification
  • Recommendation: Essential for market research and current events analysis

📚 Search Knowledge Base

  • Function: Access information from specified topics or custom knowledge bases
  • Intelligent Search Use: Core functionality for specialized domain queries
  • Best Practice: Primary tool for domain-specific expertise
  • Recommendation: Critical for all Intelligent Search Agents

📊 PPT Rendering Tool

  • Function: Generate presentations from intelligent search results
  • Enhanced Features: Document upload capabilities for PDF and Word files
  • Best Practice: Use for business intelligence reporting and academic presentations
  • Recommendation: Excellent for professional output delivery

📊 PPT Rendering Tool [Enhanced]

  • Function: Advanced presentation generation with enhanced features
  • Enhanced Capabilities: Superior formatting, advanced templates, and improved integration
  • Best Practice: Use for high-quality business presentations and detailed reports
  • Recommendation: Premium option for professional presentation needs

💻 Code Sandbox Tool

  • Function: Execute code snippets for data analysis and processing
  • Intelligent Search Use: Analyze data retrieved from knowledge sources
  • Best Practice: Use for statistical analysis and data visualization
  • Recommendation: Valuable for technical and analytical agents

🔗 MCP Services

  • Function: Automatically call MCP (Model Context Protocol) services during agent execution
  • Available Services:
    • Notion MCP: Workflow automation and productivity integration
    • Gmail MCP: Email management and communication automation
    • Expandable: "Add MCP Service" button allows integration of additional services
  • Master Toggle: Single switch to enable/disable all MCP services
  • Enhanced Features:
    • Multi-service integration in one tool
    • Seamless workflow automation across platforms
    • Expandable service ecosystem
  • Best Practice: Use for comprehensive workflow automation spanning multiple platforms
  • Recommendation: Excellent for enterprise workflow integration and productivity automation


🌐 Web Reader

  • Function: Extract and analyze content from web pages
  • Intelligent Search Use: Supplement knowledge sources with web-based information
  • Best Practice: Use for real-time content analysis and validation
  • Recommendation: Complements knowledge base searches effectively

🖼️ Image Search Tool

  • Function: Search and retrieve relevant images from various sources
  • Intelligent Search Use: Visual content to support knowledge-based findings
  • Best Practice: Use for comprehensive reports and visual presentations
  • Recommendation: Helpful for visual intelligence reports and content creation

🎨 Image Generation

  • Function: Generate custom images based on text descriptions
  • Intelligent Search Use: Create visual content to illustrate findings and concepts
  • Best Practice: Use for creating custom diagrams, illustrations, and visual aids
  • Recommendation: Excellent for creative content and visual storytelling

4.3 Tool Selection Strategy for Intelligent Search Agents

  1. Prioritize Knowledge Tools: Focus on tools that enhance knowledge source utilization
  2. Complement with Search Tools: Add tools that supplement your knowledge base
  3. Consider Output Needs: Select tools based on desired result formats
  4. Balance Functionality: Avoid tool overload - select only what you need
  5. Test Integration: Ensure tools work well together with your knowledge sources

4.4 Updated Recommended Tool Combinations

For Market Research Intelligence Agent:

  • ✅ Search Knowledge Base (core functionality)
  • ✅ Search Internet (real-time market data)
  • ✅ PPT Rendering Tool [Enhanced] (professional business reports)
  • ✅ Web Reader (competitor analysis)
  • ✅ Image Search Tool (visual market data)

For Academic Research Intelligence Agent:

  • ✅ Search Knowledge Base (academic databases)
  • ✅ Search Internet (recent publications)
  • ✅ PPT Rendering Tool (research presentations)
  • ✅ Code Sandbox Tool (data analysis)
  • ✅ Web Reader (academic content analysis)

For Business Strategy Intelligence Agent:

  • ✅ Search Knowledge Base (strategic intelligence)
  • ✅ Search Internet (market validation)
  • ✅ PPT Rendering Tool [Enhanced] (strategy presentations)
  • ✅ MCP Services (workflow integration)
  • ✅ Image Generation (custom strategy visuals)

For Business Communication Agent:

  • ✅ Search Knowledge Base (business intelligence)
  • ✅ Search Internet (current information)
  • ✅ MCP Services (Gmail + Notion integration)
  • ✅ Web Reader (content analysis)
  • ✅ PPT Rendering Tool [Enhanced] (presentation creation)

For Executive Assistant Agent:

  • ✅ Search Knowledge Base (company information)
  • ✅ MCP Services (Gmail + Notion + additional services)
  • ✅ Search Internet (real-time updates)
  • ✅ Web Reader (document analysis)
  • ✅ Image Generation (visual content creation)

For Creative Content Intelligence Agent:

  • ✅ Search Knowledge Base (content databases)
  • ✅ Search Internet (trending topics)
  • ✅ Image Search Tool (visual inspiration)
  • ✅ Image Generation (custom graphics)
  • ✅ Web Reader (content research)

4.5 Tool Selection Best Practices

Essential Tools for Most Agents:

  • Search Knowledge Base: Core functionality for all Intelligent Search Agents
  • Search Internet: Real-time information validation
  • Web Reader: Content analysis and research

Output-Focused Tools:

  • PPT Rendering Tool: Standard presentations
  • PPT Rendering Tool [Enhanced]: Professional-grade presentations
  • Image Generation: Custom visual content

Specialized Tools:

  • Code Sandbox Tool: Technical and analytical agents
  • MCP Services: Enterprise workflow integration with multiple platforms
  • Image Search Tool: Visual content research

Tool Combination Guidelines:

  1. Start with 3-4 core tools for initial agent testing
  2. Add specialized tools based on specific requirements
  3. Test tool interactions to ensure smooth operation
  4. Monitor performance - too many tools can slow response times
  5. Leverage MCP Services for comprehensive workflow automation

💾 STEP 5: Save and Deploy Your Intelligent Search Agent

5.1 Final Review for Intelligent Search Agents

Before saving, review all configurations:

  • ✅ Agent name reflects intelligent search capabilities
  • ✅ Description emphasizes knowledge source utilization
  • ✅ Custom prompt includes knowledge integration instructions
  • ✅ Appropriate tools selected for intelligent search tasks
  • ✅ Consistent formatting and professional focus maintained

5.2 Saving Your Intelligent Search Agent

  • Action: Click "Confirm" button at bottom right
  • Alternative: Click "Cancel" to abandon changes
  • Result: Your Intelligent Search Agent will be created and added to the Agent Store
  • Status: Agent will be marked as "Beta" feature

5.3 Post-Creation Next Steps

  1. Test Your Agent: Run sample queries to verify functionality
  2. Refine Performance: Adjust prompts based on initial results
  3. Monitor Usage: Track agent effectiveness and user feedback
  4. Update Regularly: Keep knowledge sources and tools current

Understanding Multi-step Search Agents

While this manual focuses on Intelligent Search Agent creation, it's important to understand the alternative method:

Multi-step Search Agent Characteristics

  • Automation: Automatically plans multiple search steps
  • Process: Breaks down complex queries into sequential searches
  • Intelligence: Uses AI to determine optimal search strategy
  • Results: Comprehensive analysis through automated planning

When to Choose Each Method

  • Intelligent Search Agent: When you need custom knowledge sources and specific expertise
  • Multi-step Search Agent: When you want automated query decomposition and planning

Best Practices for Intelligent Search Agents

Knowledge Source Optimization

  1. Define Knowledge Domains: Clearly specify your agent's knowledge areas
  2. Source Integration: Ensure prompts reference knowledge source utilization
  3. Validation Methods: Include steps for cross-referencing information
  4. Update Protocols: Plan for knowledge source maintenance

Prompt Engineering for Knowledge-Based Agents

  1. Knowledge Integration: Explicitly mention knowledge source usage
  2. Search Methodology: Define how to approach knowledge retrieval
  3. Synthesis Instructions: Specify how to combine multiple sources
  4. Quality Assurance: Include accuracy verification steps

Performance Evaluation Metrics

  1. Response Accuracy: Measure correctness of information provided
  2. Source Utilization: Track effective use of knowledge bases
  3. User Satisfaction: Monitor user feedback and engagement
  4. Task Completion: Assess successful query resolution rates

Felo Agent Store Integration

Agent Store Ecosystem

The Felo Agent Store serves as a comprehensive hub containing:

  • Intelligent Search Agents: Custom knowledge-based agents
  • Multi-step Search Agents: Automated planning agents
  • Pre-built Templates: Ready-to-use specialized agents
  • Community Contributions: User-shared agents

Categories in Agent Store

Based on current agent store content:

  • Business Intelligence: Market research, competitive analysis
  • Academic Research: Literature reviews, research synthesis
  • Content Creation: Writing assistance, topic research
  • Financial Analysis: Investment research, market analysis
  • Strategic Planning: Business model analysis, feasibility studies

Conclusion

Creating effective Felo AI Intelligent Search Agents requires careful consideration of knowledge sources, thoughtful prompt engineering, and strategic tool selection. By following this comprehensive manual and implementing consistent formatting across all examples, you can build professional AI assistants that leverage specialized knowledge to significantly enhance productivity and research capabilities.

The key to successful Intelligent Search Agent creation lies in understanding how to integrate custom knowledge sources with AI capabilities, crafting precise instructions for knowledge utilization, and selecting appropriate tools to support intelligent search processes.

The enhanced MCP Services with Gmail integration and expandable service options represents a significant advancement in workflow automation capabilities, allowing agents to seamlessly integrate with multiple business platforms and communication tools.

Remember that Intelligent Search Agent creation is currently in Beta, indicating ongoing development and feature enhancement. Stay updated with Felo AI's latest developments to maximize your agent's potential.

Whether you choose to create an Intelligent Search Agent for custom knowledge integration or utilize Multi-step Search Agents for automated query planning, the Felo Agent Store provides a comprehensive ecosystem for all your AI agent needs.

Keywords: Felo AI, Intelligent Search Agent, Multi-step Search Agent, Knowledge Base Integration, Custom AI Assistant, Agent Store, Beta Features, Business Intelligence, Academic Research, Market Analysis, MCP Services, Gmail Integration, Workflow Automation

中文摘要

Felo AI 智能搜索代理創建完整指南

本指南詳細介紹了如何使用 Felo AI 平台創建智能搜索代理（Intelligent Search Agent）。Felo AI 提供兩種代理創建方法：智能搜索代理和多步驟搜索代理，本手冊專注於智能搜索代理的創建過程。

智能搜索代理特點：

  • 使用自定義指令和特定知識源進行搜索
  • 目前處於 Beta 測試階段
  • 支持完全自定義的提示、角色和工具選擇
  • 能夠整合專業知識庫提供精準結果

標準化創建流程：

  1. 基本信息配置 - 使用一致的命名和描述格式
  2. 自定義提示編寫 - 採用標準化的四要素結構
  3. 模型選擇 - 使用 Claude 4 Sonnet 進行知識基礎搜索優化
  4. 工具管理 - 根據代理類型選擇最佳工具組合
  5. 保存部署 - 完成配置並添加到代理商店

增強的 MCP 服務：
MCP 服務現已包含 Gmail 整合功能，並支持添加更多服務，為企業工作流程自動化提供了強大的多平台整合能力。

標準化範例系統：
本指南提供了完整的標準化範例，包括市場研究智能代理、學術研究智能代理、商業策略智能代理、商業溝通代理和執行助理代理，每個範例都遵循相同的結構和專業水準。

最佳實踐：

  • 重視知識源整合和利用
  • 採用標準化的提示工程方法
  • 選擇支持智能搜索的工具組合
  • 實施性能評估和持續優化
  • 善用 MCP 服務進行工作流程自動化

通過實施一致的格式化方法和專業導向的範例，用戶可以創建高質量、標準化的智能搜索代理，有效提升工作效率和研究能力。

— Dr. Ken FONG


Keywords: Felo AI, Intelligent Search Agent, Multi-step Search Agent, Knowledge Base Integration, Custom AI Assistant, Agent Store, Beta Features, Business Intelligence, Academic Research, Market Analysis, MCP Services, Gmail Integration, Workflow Automation, 智能搜索代理, 知識庫整合, 代理商店, 商業智能, 學術研究, 工作流程自動化

Tags
Ken Fong

Posted by Ken Fong

Post a Comment

0 Comments