Agents Overview
Agents are the core of 88agents. Each agent is an AI-powered assistant configured for specific tasks, trained on your data, and deployed to interact with your users.
What is an Agent?
An agent is a complete AI assistant package that includes:
- Identity — Name, avatar, and branding
- Personality — System prompt, tone, and behavior guidelines
- Intelligence — LLM model selection and configuration
- Knowledge — Training data and information sources
- Channels — Deployment endpoints (widget, Slack, API)
Each agent operates independently with its own settings and knowledge base. You can create unlimited agents within a workspace to handle different use cases.
Agent Components
| Component | Description | Learn More |
|---|---|---|
| Personality | System prompt, tone, response style | Personality Guide |
| LLM Settings | Model selection, temperature, max tokens | LLM Configuration |
| Knowledge | Documents, URLs, text sources | Knowledge Sources |
| Channels | Widget, Slack, API endpoints | Channels Guide |
| Monitoring | Conversations, logs, analytics | Monitoring Guide |
| Manifest | Declarative config defining the full agent shape | Manifest Reference |
Best Practices
- One agent per use case — Create separate agents for support, sales, and FAQ rather than one do-everything agent
- Start focused — Begin with a narrow knowledge base and expand as needed
- Test thoroughly — Use the playground to test edge cases before deploying
- Monitor conversations — Review chat logs regularly to identify improvement areas
- Iterate quickly — Don't wait for perfection; deploy early and refine based on real usage. You can revisit the Personality and Tuning pages to fine tune the System Prompt and adjust how many knowledge sources are retrieved.