Skip to main content

​
What are Outcomes?

Outcomes allow you to extract structured, formatted data from your AI agent conversations. Define a schema that specifies exactly what information to capture, and your agent will automatically extract and return that data in a consistent JSON format after each call, enabling seamless integration with your CRM, databases, and business systems.
Key Features:
  • Structured Data Extraction - Convert free-form conversations to JSON
  • Custom Schemas - Define any data structure you need
  • Type Safety - Support for strings, numbers, booleans, objects, and arrays
  • Visual Builder - Create schemas without writing code
  • JSON Editor - Advanced users can edit raw JSON Schema
  • Real-time Validation - Catch errors before saving

​
How It Works

The outcome system transforms unstructured conversations into structured data:
  1. Define Schema - Specify what data to extract (name, email, intent, etc.)
  2. Configure Agent - Add the schema to your agent’s outcome settings
  3. Conduct Call - Agent has a conversation with the user
  4. Extract Data - AI analyzes the conversation against your schema
  5. Return Results - Structured data sent via webhook or available in call history

​
Supported Data Types

Primitive Types
  • String - Text data, names, descriptions, notes
  • Number - Quantities, amounts, ratings, scores
  • Boolean - Yes/no questions, flags, binary decisions
Complex Types
  • String with Enum - Predefined categories and classifications
  • Object - Nested structures for grouped related data
  • Array - Lists of items, collections, repeated data

​
Common Use Cases

Sales Call Qualification
  • Extract lead information and company details
  • Capture budget ranges and purchase timelines
  • Identify pain points and qualification status
Appointment Scheduling
  • Gather patient/customer information
  • Capture preferred dates and time slots
  • Record appointment types and requirements
Customer Feedback Collection
  • Collect satisfaction ratings and scores
  • Identify positive and negative aspects
  • Capture additional comments and follow-up preferences
Order Processing
  • Extract customer and shipping information
  • Capture ordered items with quantities and prices
  • Record payment methods and order totals

​
Getting Started

Choose your implementation path based on how you want to work with outcomes:

Configure in Dashboard

Build outcome schemas using the visual builder or JSON editor. Perfect for defining data structures and testing extraction logic.

Access via API

Retrieve outcome data programmatically from completed calls. Ideal for integrating extracted data with your systems.

​
Learn More

Webhooks

Receive outcome data in real-time

Call History

View outcome results in the dashboard

Single Prompt Agents

Configure outcomes for Single Prompt Agents

Flow Agents

Set up outcomes in Flow Agents

​
Best Practices

Schema Design
  • Use clear, consistent field names (snake_case recommended)
  • Write helpful descriptions that guide the AI
  • Only mark fields as required if they’re essential
  • Group related fields using object types
Data Quality
  • Keep schemas focused on specific outcomes
  • Use enums for predefined categories
  • Choose appropriate data types (number vs string)
  • Test schemas with real conversations
Integration
  • Access outcome data via webhooks for real-time processing
  • Store extracted data in your CRM or database
  • Use outcomes to trigger business workflows
  • Monitor extraction quality and adjust schemas as needed
Validation
  • Test schemas before deploying to production
  • Review call transcripts if data is missing or incorrect
  • Ensure agents ask for required information during calls
  • Adjust prompts to gather necessary data