Let's review
In this module, you’ve learned about the AI Agent action in Tines. You've learned how to configure task mode for processing data, and chat mode for conversational experiences. You've explored how to attach various types of tools, structure outputs with schemas, and monitor credit usage to build effective, cost-conscious AI agents.
Key points to remember:
Task mode uses system instructions and prompts to process data and return results, while chat mode uses system instructions and an initial message to create conversational experiences with clear completion goals defined in the system instructions.
Tools (templates, Send to Story, custom tools, internal tools, and MCP servers) transform AI agents into autonomous problem-solvers. Design for specific tool usage patterns like sequential, conditional, or parallel execution, and provide clear tool descriptions to help the AI decide when and how to use them.
Specialized agents outperform generalist agents. Focus each agent on a specific task, use output schemas for consistency, monitor credit usage with alerts, and refine your agents based on conversation histories to improve reliability and reduce costs.
Next up
In the next module, you'll explore MCP servers and learn how to expose Tines tools to external AI applications using the Model Context Protocol. This opens up new possibilities for integrating Tines with AI-powered tools and platforms.