Microsoft Certified: Azure AI Engineer AssociatePlan and manage an Azure AI solutionMedium

A company is planning to deploy an Azure AI solution that includes several Azure Cognitive Services, such as Text Analytics, Speech, and Computer Vision. The solution needs to be able to scale independently for each service and maintain separate billing and access control for different departments. Which type of Cognitive Services resource should you provision?

  1. AAzure Bot Service with Cognitive Services integration.
  2. BIndividual Cognitive Services resources for each service.
  3. CAn Azure Machine Learning workspace with integrated Cognitive Services.
  4. DA multi-service Cognitive Services resource.
Show answer & explanation

Correct answer: B. Individual Cognitive Services resources for each service.

Provisioning individual Cognitive Services resources allows for independent scaling, separate billing, and granular access control (via RBAC) for each service. This directly addresses the requirement for independent management per department and service.

Why the other options are wrong

  • A. Azure Bot Service is for building conversational AI experiences. While it can integrate Cognitive Services, it's not the primary resource for provisioning and managing the underlying Cognitive Services themselves for independent scaling and billing.
  • C. An Azure Machine Learning workspace is for machine learning model development and deployment, not for provisioning and managing Cognitive Services instances for independent scaling and billing.
  • D. A multi-service resource (aka 'Cognitive Services' resource) provides access to multiple services under a single key/endpoint, but it doesn't allow for separate billing, independent scaling, or granular access control per individual service.

Cognitive Services Resource Types

Azure Cognitive Services can be provisioned as either a single-service resource (for specific service like Text Analytics) or a multi-service resource (for access to multiple services with a single key/endpoint).

  • Single-service resources offer granular control and billing.
  • Multi-service resources simplify management for combined usage.
  • Choice depends on specific requirements for isolation and billing.

Memory trick: Separate services, separate bills, separate control: Think 'individual' not 'bundle'.

More Plan and manage an Azure AI solution questions