Dialogflow CX - data store response schema

Google Cloud Build with AI Conversational Agents (Dialogflow CX)
1

Hi Team,

I was trying the find the exact schema structure for data store responses. I have a data store that’s been configured with FAQ and unstructured documents. By default, in the agent responses, we get “$request.knowledge.answers[0]”. I logged the response structure for “$request.knowledge” and found response is something similar to below. I wanted to confirm with the documentation to ensure is that exactly same and what all other key value pairs exist. But I couldn’t find the documentation for this schema. Can you please help ? Reason is, I found for FAQ documents => [details.response_reason=faq] and for Unstructured documents => [details.response_reason=normal_execution]. So I wanted to re-confirm this with the documentation, so I can do some transitions in my flow based on this values.

Schema

{
“answers”: [
“<answer_1>”
],
“questions”: [
“<question_1>”
],
“sources”: [
[
{
“documentId”: “<docuement_id>”,
“title”: “”
}
]
],
“grounding_confidences”: ,
“details”: {
“response_reason”: “faq”
}
}

1 Like
2

I cannot confirm it 100% since I couldn’t find docs referencing it but so far I’ve seen the same schema!

Best,

Xavi

4 Likes
3

Thanks @xavidop . I have tested multiple data store responses for both faq and unstructured documents. There’s no changes found in Schema. For FAQ, response_reason was ‘faq’ and grounding_confidences was . Similarly for unstructured docs, response_reason was ‘normal_execution’ and grounding_confidences was [“<confidence_level>”]. Based on these, i’ve made my transition logics.

can this something be raised as a request in issue tracker?

5 Likes
4

yeah, you can submit product request via issue tracker!

4 Likes