GenerativeAgent uses LLMs and other generative AI technology, allowing for human-like interactions and reasoning but also requiring careful consideration on the instructions and functions you define to ensure it behaves as expected.

Creating successful tasks is an iterative process. We have multiple resources and tools to help improve task performance:

Task Best Practices

A list of different strategies and approaches to improve task performance.

Conditional Templates

Use conditional logic to dynamically change the instructions shown to GenerativeAgent.

Trial Mode

Use Trial Mode to test whether GenerativeAgent can use new Functions correctly before full rolling them out in production.

Keep Fields

Use Keep Fields to keep limit the data saved when calling a function.

Was this page helpful?

Deploying to GenerativeAgentTasks Best Practices