What is an Agent?
In the Connect platform, an "agent" refers to an AI-powered conversational interface that can understand natural language, maintain context, and provide helpful responses based on both general knowledge and your specific information.
Agents vs. Traditional Chatbots
While many people use the terms "agent" and "chatbot" interchangeably, there are significant differences between Connect agents and traditional chatbots:
| Feature | Traditional Chatbots | Connect Agents |
|---|---|---|
| Understanding | Rule-based pattern matching | Natural language understanding |
| Knowledge | Pre-programmed responses | Access to large language models + your knowledge |
| Context | Limited or none | Maintains conversation history and context |
| Learning | Manual updates required | Improves through usage and feedback |
| Conversations | Follows predefined paths | Dynamic and natural interactions |
| Complexity | Best for simple, predictable tasks | Handles nuance and complexity |
| Setup | Requires extensive rule programming | Configuration over programming |
| Maintenance | Requires constant rule updates | Knowledge-focused updates |
Key Components of a Connect Agent
A Connect agent consists of several key components:
1. Language Model Foundation
Connect agents are built on advanced large language models (LLMs) that provide:
- Natural language understanding
- Context awareness
- Reasoning capabilities
- Response generation
2. Knowledge Integration
Agents can connect to various knowledge sources:
- Your organization's documents
- Databases
- APIs
- Web content
- Structured data
3. Configuration and Personality
Each agent has configurable parameters including:
- Tone and style
- Response length preferences
- Creativity vs. accuracy balance
- Specialized behavior instructions
4. Integration Capabilities
Agents can be deployed across multiple channels:
- Web applications
- Mobile apps
- Messaging platforms
- Custom interfaces
When to Use Connect Agents
Connect agents are ideal for scenarios where:
- Conversations are unpredictable: Users might ask questions in various ways
- Complex information is involved: The agent needs to understand and explain sophisticated concepts
- Context matters: Previous parts of the conversation influence current responses
- Personalization is important: The agent should adapt to individual users
- Knowledge evolves: Information needs regular updating without rebuilding the entire system
Common Agent Types
Knowledge Assistants
Specialized in retrieving and explaining information from your knowledge base.
Example use cases:
- Internal documentation search
- Product support
- Policy clarification
Process Guides
Walk users through multi-step processes with contextual awareness.
Example use cases:
- Onboarding flows
- Form completion assistance
- Technical troubleshooting
Conversational Interfaces
Provide natural dialogue experiences for your applications.
Example use cases:
- Customer service
- Virtual assistants
- Interactive help systems
Building Your First Agent
Ready to create your own agent? Follow these steps:
- Define your agent's purpose and scope
- Gather the knowledge it will need access to
- Configure agent settings and personality
- Test with realistic scenarios
- Deploy to your chosen channels
- Monitor performance and gather feedback
- Continuously improve based on usage data
For step-by-step instructions, continue to our guide on creating agents.