Skip to main content

Post-Process

A Post-Process Skill is able to observe or change a Skill's output before it is sent to the Digital Person to be turned into speech.

The Skill should handle an ExecuteResponse, observe it, optionally modify it, and then respond with the same or modified ExecuteResponse to be sent back to the Digital Person for speech processing.

Common use cases​

A Post-Process Skill may observe or modify an ExecuteResponse, which enables the following use cases:

  • Data Collection to analyse Skill response data and a Digital Person's spoken responses.
  • Response Augmentation to add any additional markup, context or data to the response before final processing.

Implementation requirements​

Implementation requirements for a Post-Process Skill:

  • It must consider speed, as it will be executed synchronously for every Skill response.
  • It must respond with a valid ExecuteResponse.

Creating a Post-Process Skill​

Post-Process Skills must be created programmatically using the Skill SDK.

Registering a Post-Process Skill​

skill-definition.json
{
  "name": "My Skill",
  "summary": "skill summary here",
  "description": "",
  // ... other skill properties
  "config": {
    "skillType": "POST_PROCESS"
  }
}