Business Card Reader AI Agent
Overview
The Business Card Reader helps you capture contact information from a business card image using Document AI (OCR + data extraction). It converts the card into structured contact data so you can quickly create or update contacts without manual typing.
Step 1: Scan a business card using Document AI Agent
Go to the AI Agent Assistants library.
Select OCR & Data Extraction from the category list.
Step 2: Use the Demo tab to test
Open the Demo tab.
Upload an image/document in the chat area (paperclip/attachment option).
Review the extracted output returned by the agent.
Step 3: Configure agent settings (optional)
Use the top tabs to configure the agent based on your needs:
Basic Information: update agent name/description and display details.
Behavior Settings: control how the agent responds, summarizes extracted data, and asks follow-up questions when information is missing.
Knowledge Support: add guidance, rules, or reference content to improve consistency.
Multi Agent: use when the Document AI agent needs to work together with other agents.
Advanced Settings: fine-tuning and technical options for administrators.
Click Save after any changes.
Step 4: Add workflow-based Functions (Capabilities) (optional)
Under Behavior Settings, in Capabilities, you can enhance the agent using workflow-based functions so it can take actions.
Click + Add Functions to attach a function/workflow.
Attached functions appear under List of Functions.
Use the delete icon to remove a function from the agent (this removes access from the agent; it does not necessarily delete the original workflow).
Best practice: keep function names clear and specific, and only add what the agent actually needs.
Input quality guidelines (recommended)
For best OCR/extraction accuracy:
Use clear, sharp images (avoid blur)
Ensure good lighting (avoid shadows and glare)
Keep the document upright where possible
Avoid heavily compressed or pixelated images
Last updated