Introduction to AI agents

make ai agents is available in beta to all plans using make's ai provider, with the option to use your own connection on pro plans and higher as a beta feature, its functionality and pricing may change an ai agent is an autonomous system that acts to achieve its goals while following certain r ules and conditions think of it like a skilled assistant who performs tasks for you after understanding what you want to do {{product name}} ai agents consist of a few fundamental components that can be thought of as a brain (llm), instructions (system prompt), memory (context), and tools (scenarios and mcp tools) a large language model (llm) acts as the agent's brain, powering its reasoning for decision making the system prompt instructs the agent on how to behave like a job description, it defines its role, limitations, and processes context, or files stored in the agent's long term memory, provides extra information that helps the agent tailor its responses to your needs for example, an internal knowledge base or company policies t ools provide the agent with the capabilities to do its job the agent runs tools for you to complete your requested tasks in make, the agent's tools can include {{scenario plural lowercase}} and mcp tools {{scenario plural}} used as tools for agents must be scheduled on demand or triggered by a custom webhook the agent uses tool names and descriptions to know which tool to use to achieve a goal it uses scenario inputs and outputs docid 7odicud0ke9 tr4xwuiqn names and descriptions to understand the data it receives or sends back to you in the next section, you will explore these concepts in more detail the make ai agent's reasoning is transparent and easy to review however, llms are still evolving, which could impact agent outcomes always check outputs to ensure their accuracy and reliability