Agent Intelligence Overview

Agent Intelligence features enable your AI agents to remember important information, store reusable outputs, and provide personalized, context-aware responses over time.

Why Agent Intelligence Matters

Without memory and context features, every agent interaction starts from scratch. Agents would:

• Ask the same questions repeatedly
• Forget user preferences and history
• Redo research that was already completed
• Miss important organizational policies

Agent Intelligence solves these problems by giving agents the ability to:

1. Remember - Store facts about contacts, preferences, and organizational policies
2. Reuse - Cache expensive research and analysis for future use
3. Learn - Build up knowledge over time from interactions

Components

Memories

Memories store facts and context that agents can recall during processing:

Memory Type	Description	Availability
Personal Memories	Facts about individual contacts (preferences, projects, communication style)	All plans
Global Memories	Company-wide policies and guidelines injected into all agent executions	Pro Plus

Example use cases:

• Personal: "John prefers email responses in bullet points"
• Personal: "Sarah is working on Project Phoenix with deadline March 15"
• Global: "Always mention our 30-day return policy for product inquiries"
• Global: "Escalate any emails mentioning legal action to the legal team"

Artifacts

Artifacts store large agent-generated outputs for reuse:

• Research reports
• Data analyses
• Processed datasets
• Generated documents

Key benefits:

• Avoid redundant work - Don't repeat expensive research
• Reduce context bloat - Reference artifacts instead of including full content
• Enable sharing - Multiple agents can access the same artifact
• Track usage - See how often artifacts are accessed

How It Works

Memory Flow

Email Received
    |
    v
Agent Loads Context
    |---> Global Memories (policies, guidelines)
    |---> Personal Memories (sender-specific facts)
    |
    v
Agent Processes Email
    |
    v
Agent May Store New Memories
    |---> "Contact prefers formal tone"
    |---> "Mentioned budget of $50k for Q2"
    |
    v
Response Sent

Artifact Flow

Research Request Received
    |
    v
Agent Searches Existing Artifacts
    |
    |---> Cache Hit? Use existing artifact
    |
    v (if no match)
Agent Performs Research
    |
    v
Agent Stores Artifact
    |---> Title, Summary, Content
    |---> Tags, Category
    |---> Validity Period
    |
    v
Future Requests Reuse Artifact

Automatic vs Manual Management

Agent-Created

Agents can create memories and artifacts automatically during execution:

• Memory tools: manage_personal_memory, manage_global_memory
• Artifact tools: store_artifact

Configure agents with appropriate instructions to use these tools effectively.

Admin-Created

Administrators can manually create and manage:

• Global memories - Define organizational policies
• Personal memories - Pre-populate contact information

Access these through:

• Admin > Data & Logs > Memories for both memory types
• Admin > Data & Logs > Artifacts for artifact management

Best Practices

For Memory Usage

1. Start with global policies - Create global memories for organization-wide rules before deploying agents
2. Let agents learn - Allow agents to create personal memories from interactions
3. Review regularly - Periodically audit memories for accuracy and relevance
4. Set expirations - Use expiration dates for time-sensitive information

For Artifact Usage

1. Enable for research agents - Assign artifact tools to agents that perform expensive research
2. Use appropriate TTLs - Set validity periods based on content freshness needs
3. Categorize consistently - Use consistent categories and tags for discoverability
4. Monitor usage - Review artifact access patterns to optimize caching

Feature Availability

Feature	Starter	Pro	Pro Plus
Personal Memories	Yes	Yes	Yes
Global Memories	-	-	Yes
Artifacts	Yes	Yes	Yes

Getting Started

1. Set up Personal Memories - Learn how agents store contact-specific information
2. Configure Global Memories - Define organizational policies (Pro Plus)
3. Use Artifacts - Enable agents to cache and reuse research

Related Topics

• Tools - Memory and artifact tool descriptions
• Agents - Agent configuration
• Agent Executions - View memory and artifact usage in executions