The platform returns a strict 400 error when attempting to provision a new chatbot via the Platform API. The response body is minimal: {"code":"bad_request","message":"Invalid request body."}. No specific field validation errors are included in the response, making debugging difficult.
The application is a Premium App listed on AppFoundry, using multi-org OAuth 2.0 with a service account. The request targets the POST /api/v2/ai/bots/chatbots endpoint. The payload matches the schema defined in the latest OpenAPI spec for Genesys Cloud version 2023-10-01. All required fields, including name, type, and knowledge_base_ids, are present and formatted correctly.
We have verified the service account permissions. The role includes bot:bot:view and bot:bot:create. The issue persists across three different tenant environments. Standard cURL tests with the same JSON payload succeed when executed directly from a browser-based developer console, suggesting the issue might be related to how the SDK serializes the request or specific header requirements for AI resources.
The integration uses the Python SDK v4.2.1. Logs show the request is signed correctly, and the authorization token is valid. Has anyone encountered silent failures with the chatbot provisioning endpoint? Any insights into hidden validation rules or SDK serialization quirks for AI/Bot resources would be appreciated.