Automating repetitive actions and orchestrating multi-step workflows with tool and API calls is correct because Agents for Amazon Bedrock are designed to decompose user intent into steps and invoke external tools and APIs to complete tasks, which directly supports handling high volumes of support tickets.

Automating repetitive actions and orchestrating multi-step workflows with tool and API calls lets agents plan sequences of actions use action groups to call services retrieve relevant knowledge and coordinate end to end resolutions so workflows are automated and human effort is reduced, which helps scale to thousands of daily tickets.

Automatically building and training new proprietary foundation models to anticipate customer intent is incorrect because agents do not train or create foundation models. Building and training models is done with separate services such as Amazon SageMaker and is not an inherent agent capability.

Offloading all task coordination and state management to AWS Step Functions is incorrect because AWS Step Functions is a separate orchestration service. Agents can call Step Functions as a tool but they do not automatically transfer all coordination and state handling to it.

Choosing and switching between multiple foundation models and merging their outputs without setup is incorrect because agents do not automatically ensemble or route between models without explicit configuration. Model selection and output merging require the builder to configure routing or ensembling.

Full AWS Practitioner Certification Question