Quick Answer: What is the AI Agent Node?

[!NOTE] The AI Agent Node is an autonomous processing block that uses advanced Generative AI (LLMs) to perform tasks that require human-like understanding. Unlike static conversion nodes, the AI Agent can summarize text, extract complex patterns, generate new content, and make logical "if-this-then-that" decisions based on the content of your documents.

Core Capabilities

1. Autonomous Summarization

Feed a 50-page legal PDF into the AI Agent and use a prompt like "Summarize the key termination clauses into 3 bullet points." The node will output the synthesis, ready for email delivery or storage.

2. Logical Gating

The AI Agent can act as a sophisticated router. For example: "Analyze this invoice. If the total is over $1,000, output 'high'; otherwise output 'low'." You can then connect a downstream Switch Node to handle the results differently.

3. Data Transformation

Convert unstructured OCR text into clean Markdown or JSON without writing regular expressions.

Configuration Guide

Setting	Description	Best Practice
System Prompt	The overarching instructions for the AI.	Use "Act as a legal assistant" to set the persona.
User Prompt	The specific instruction for this execution.	Be specific: "Extract only the date and amount."
Temperature	Controls "creativity" or randomness.	Use `0` for extraction/logic; use `0.7` for creative writing.
Credits	The platform cost per execution.	High-capacity models consume more credits than base models.

Best Practices

Context is King: Always provide enough surrounding text to ensure the AI doesn't hallucinate.
Structured Outputs: Ask the AI to "Return only valid JSON" if you plan to use the data in a Spreadsheet or Database node.
Error Handling: Use the Execution History panel to debug prompts if the AI returns unexpected results.

[!TIP] Need to extract specific table data? Use the specialized AI Extract Node instead for higher precision on structured data sets.