Skip to main content

​
Overview

The Search Chat History tool allows you to search through all your previous conversations and uploaded documents to find relevant information. This powerful tool is essential for building knowledge continuity, tracking agent performance, and creating intelligent monitoring systems.

​
Key Capabilities

  • Full-text search across all chat messages and AI responses
  • Semantic matching to find conceptually related content
  • Time-based filtering with recent vs relevant sorting
  • Context preservation across multiple conversation threads
  • File content search across all uploaded documents
  • Multi-format support including PDFs, Word docs, spreadsheets, and images
  • Metadata search including file names, creation dates, and tags
  • Cross-document correlation to find related information

​
Agent Performance Tracking

  • Execution monitoring for scheduled and recurring agents
  • Performance metrics including success rates, execution times, and resource usage
  • Pattern recognition to identify trends and anomalies
  • KPI aggregation for building performance dashboards

​
How It Works

Simply describe what you’re looking for:
  • “Find information about Q3 sales metrics”
  • “Show me discussions about the new product launch”
  • “Search for customer feedback on shipping times”

​
Advanced Search Patterns

Use specific queries for targeted results:
  • Agent-specific: “daily-sales-collector revenue metrics”
  • Tagged content: “#weekly-report performance data”
  • Time-based: “error-handler failures last week”
  • Pattern matching: “verification required approved rejected”

​
Search Parameters

  • Query: The main search terms (required)
  • Sort: Choose between “recent” (chronological) or “relevant” (semantic similarity)
  • Chat ID: Search scope control:
    • Leave empty (default) to search across ALL your chat history and documents
    • Provide a specific chat ID to search within one conversation only
  • Scope: Search across all conversations or specific timeframes

​
Use Cases

​
Knowledge Management

Build institutional knowledge by searching past conversations:
  • Decision tracking: Find rationale behind past decisions
  • Process documentation: Locate established procedures and workflows
  • Expertise location: Identify who discussed specific topics
  • Learning continuity: Build on previous research and analysis

​
Agent Monitoring

Track and optimize agent performance:
  • Success rate monitoring: Identify which agents perform best
  • Error pattern analysis: Find common failure points
  • Resource optimization: Track token usage and execution times
  • Performance benchmarking: Compare agent performance over time

​
KPI Development

Build comprehensive monitoring systems:
  • Metrics aggregation: Collect KPIs from multiple agents
  • Trend analysis: Identify patterns in performance data
  • Alert systems: Monitor for performance degradation
  • Dashboard creation: Build executive reporting systems

​
Micro-Agent Integration

​
Performance Monitoring Agents

Create agents that monitor other agents: 
Performance-Monitor-Agent:
  Schedule: Daily
  Steps:
    1. Search: "daily-sales-collector execution metrics"
    2. Calculate success rate and average execution time
    3. Generate performance report
    4. Alert if performance drops below threshold

​
KPI Aggregation Agents

Build agents that compile metrics from multiple sources: 
KPI-Compiler-Agent:
  Schedule: Weekly
  Steps:
    1. Search: "sales-tracker revenue metrics"
    2. Search: "customer-service satisfaction scores"
    3. Search: "marketing-agent campaign performance"
    4. Compile comprehensive business dashboard
    5. Email to executive team

​
Self-Improving Networks

Create agents that optimize other agents: 
Agent-Optimizer:
  Schedule: Monthly
  Steps:
    1. Search: "all agent execution results"
    2. Identify underperforming agents
    3. Analyze failure patterns
    4. Generate optimization recommendations
    5. Update agent configurations

​
Best Practices

​
Query Construction

  • Be specific: Use exact agent names and clear terms
  • Use consistent terminology: Establish standard terms across agents
  • Include context: Add relevant timeframes and scope
  • Try multiple phrasings: Different wording may yield different results

​
Data Organization

  • Use tags: Add hashtags to agent outputs for easy filtering
  • Consistent naming: Use descriptive, consistent names for agents
  • Time markers: Include timestamps and date references
  • Structured output: Format agent results consistently

​
Performance Optimization

  • Targeted searches: Use specific queries rather than broad searches
  • Time-based filtering: Limit search scope when possible
  • Result analysis: Review search results to refine queries
  • Caching strategies: Store frequently accessed results

​
Integration with Other Tools

​
Agent Creation

Use search results to inform new agent creation:
  • Pattern identification: Find recurring tasks that need automation
  • Template creation: Build agents based on successful patterns
  • Workflow optimization: Identify bottlenecks and improvement opportunities

​
Document Management

Combine with document tools for comprehensive knowledge management:
  • Research compilation: Gather information from multiple sources
  • Report generation: Create summaries based on historical data
  • Knowledge base building: Organize information for future reference

​
Communication Tools

Integrate with email and messaging for automated reporting:
  • Status updates: Send regular performance reports
  • Alert systems: Notify stakeholders of important findings
  • Dashboard distribution: Share KPI reports with teams

​
Limitations

​
Search Scope

  • Conversation-specific: Only searches within your organization’s conversations
  • Time boundaries: Limited to available conversation history
  • Access permissions: Respects user permissions and privacy settings

​
Content Types

  • Text-based: Primarily searches text content in messages and documents
  • Language support: Best performance with English content
  • File format limitations: Some specialized formats may have limited search capability

​
Performance Considerations

  • Large result sets: Very broad searches may return overwhelming results
  • Processing time: Complex searches may take longer to complete
  • Resource usage: Frequent searches may impact system performance

​
Troubleshooting

​
No Results Found

  • Check spelling: Verify search terms are correct
  • Try synonyms: Use alternative terms for the same concept
  • Broaden scope: Remove overly specific filters
  • Check timeframe: Ensure you’re searching the right time period

​
Too Many Results

  • Add specificity: Include more specific terms or agent names
  • Use time filters: Limit search to recent conversations
  • Try exact phrases: Use quotes for exact phrase matching
  • Filter by tags: Use hashtags to narrow results

​
Irrelevant Results

  • Refine query: Use more specific or technical terms
  • Check context: Ensure search terms match your intent
  • Use recent sort: Switch to chronological sorting
  • Try different phrasing: Rephrase the query with different words
Need help setting up search-based monitoring? Check our agent examples or contact support.
⌘I