Skip to main content

​
Skills & Tool Calls

Skills are custom JavaScript functions that extend your AI agent’s capabilities. When deployed, they become available as tools that the AI can invoke during conversations.

​
Architecture

​
Tool Call Flow

  1. Tool Discovery: AI agent analyzes available skills and their parameters
  2. Parameter Extraction: Agent extracts relevant data from conversation context
  3. Skill Invocation: Platform executes the skill with provided parameters
  4. Response Processing: Output is validated and formatted for the user
  5. Context Integration: Results are integrated back into the conversation

​
When Skills Are Called

  • User requests specific actions (“book a meeting”, “check inventory”)
  • AI determines contextual actions are needed
  • Explicit button/carousel interactions trigger specific skills
  • Scheduled or event-driven executions (future capability)

​
Skill Types

​
Informational Skills

Fetch and present data without side effects. Examples: Weather lookup, stock prices, knowledge base search

​
Action Skills

Perform operations that change external state. Examples: Create calendar events, send emails, update CRM records

​
Interactive Skills

Present UI elements for continued user interaction. Examples: Multi-step forms, approval workflows, guided tutorials

​
Tool Call Context

When a skill executes, it receives rich context about the conversation:
  • Conversation Metadata: ID, participant info, channel type
  • Organization Context: Settings, timezone, custom fields
  • Message History: Recent messages for context-aware responses
  • User Profile: Contact information and preferences
This context enables skills to provide personalized, relevant responses.

Developement