Need help with dynamic AI content generation
I'm building a chatbot that needs to:
Collect business info from conversation (
working)
Have an AI agent generate personalized content based on that info ( working)
Set individual variables with that generated content for API submission (
struggling)
Current Issue:
My agent generates great personalized content, but I can't figure out how to capture that generated content and set it to specific variables.
What I've Tried:
Set Variable cards with prompts → they generate conversational responses instead of clean content
Agent instructions to "set variables" → agent doesn't actually set them
Using {last_agent_response} → contains all content mixed together, need individual fields
JSON parsing approaches → various webhook/API integration failures
What I Need:
A reliable way to take AI-generated content (like a title the agent creates) and set it to a specific variable (like {title}) so my API tool can send structured data to a webhook.
Question:
Is there a recommended pattern in Voiceflow for capturing dynamic AI agent outputs and mapping them to individual variables? Or am I approaching this wrong?
Any guidance would be much appreciated! Been troubleshooting this for hours and hitting walls with the current approaches.
Collect business info from conversation (
working)Have an AI agent generate personalized content based on that info (
working)Set individual variables with that generated content for API submission (
struggling)Current Issue:
My agent generates great personalized content, but I can't figure out how to capture that generated content and set it to specific variables.
What I've Tried:
Set Variable cards with prompts → they generate conversational responses instead of clean content
Agent instructions to "set variables" → agent doesn't actually set them
Using {last_agent_response} → contains all content mixed together, need individual fields
JSON parsing approaches → various webhook/API integration failures
What I Need:
A reliable way to take AI-generated content (like a title the agent creates) and set it to a specific variable (like {title}) so my API tool can send structured data to a webhook.
Question:
Is there a recommended pattern in Voiceflow for capturing dynamic AI agent outputs and mapping them to individual variables? Or am I approaching this wrong?
Any guidance would be much appreciated! Been troubleshooting this for hours and hitting walls with the current approaches.
