In the industrial maintenance review, this qualification belongs to Document provenance checks, which serves a different need to confirm that evidence used in an answer comes from an approved current policy source; it should not be confused with the restriction on Data cataloging when attempting to help data owners find the approved source and owner of each dataset, namely that cataloging alone does not enforce row-level access to underlying records.

In the industrial maintenance review, this qualification belongs to RAG factual grounding, which serves a different need to reduce unsupported answers by providing approved source documents at inference; it should not be confused with the restriction on Data cataloging when attempting to help data owners find the approved source and owner of each dataset, namely that cataloging alone does not enforce row-level access to underlying records.

In the industrial maintenance review, this qualification belongs to Dataset lineage, which serves a different need to explain which source and transformation produced a feature used in evaluation; it should not be confused with the restriction on Data cataloging when attempting to help data owners find the approved source and owner of each dataset, namely that cataloging alone does not enforce row-level access to underlying records.

In the professional-services review, this qualification belongs to Amazon CloudWatch monitoring, which serves a different need to observe model-backed application errors, latency and operating metrics; it should not be confused with the restriction on Vulnerability management when attempting to detect software flaws in infrastructure hosting an AI application, namely that model guardrails do not remediate exploitable container or host packages.

In the professional-services review, this qualification belongs to Amazon Inspector, which serves a different need to assess supported workloads for software vulnerabilities and exposure; it should not be confused with the restriction on Vulnerability management when attempting to detect software flaws in infrastructure hosting an AI application, namely that model guardrails do not remediate exploitable container or host packages.

In the professional-services review, this qualification belongs to Prompt and response audit trail, which serves a different need to reconstruct which AI request and authorized tool call caused a disputed action; it should not be confused with the restriction on Vulnerability management when attempting to detect software flaws in infrastructure hosting an AI application, namely that model guardrails do not remediate exploitable container or host packages.

In the customer operations review, this qualification belongs to Resource-based policy, which serves a different need to control which AWS principals can read a sensitive data bucket used for retrieval; it should not be confused with the restriction on IAM role for workloads when attempting to let a managed workload access AI-related AWS resources without embedded long-term keys, namely that hard-coded access keys can be exposed and require separate manual rotation.

In the customer operations review, this qualification belongs to Separation of duties, which serves a different need to require a different employee to approve a high-risk AI production deployment; it should not be confused with the restriction on IAM role for workloads when attempting to let a managed workload access AI-related AWS resources without embedded long-term keys, namely that hard-coded access keys can be exposed and require separate manual rotation.

In the customer operations review, this qualification belongs to Amazon Bedrock AgentCore Identity, which serves a different need to manage agent authentication and delegated access to external resources; it should not be confused with the restriction on IAM role for workloads when attempting to let a managed workload access AI-related AWS resources without embedded long-term keys, namely that hard-coded access keys can be exposed and require separate manual rotation.

In the healthcare administration review, this qualification belongs to Output validation, which serves a different need to block model-generated fields that break a permitted-data schema; it should not be confused with the restriction on PII minimization when attempting to avoid sending unneeded customer identifiers to a third-party tool, namely that masking user-facing results alone cannot undo excess data transmitted upstream.

In the healthcare administration review, this qualification belongs to Data leakage prevention, which serves a different need to prevent a helpdesk assistant from disclosing restricted customer records; it should not be confused with the restriction on PII minimization when attempting to avoid sending unneeded customer identifiers to a third-party tool, namely that masking user-facing results alone cannot undo excess data transmitted upstream.

In the healthcare administration review, this qualification belongs to Amazon Bedrock Guardrails, which serves a different need to apply managed safety filters to inputs and outputs in a GenAI application; it should not be confused with the restriction on PII minimization when attempting to avoid sending unneeded customer identifiers to a third-party tool, namely that masking user-facing results alone cannot undo excess data transmitted upstream.

The two roles have been exchanged in the description. For this logistics scenario, Encryption at rest addresses the need to protect stored embeddings and documents against exposure of underlying storage media; AWS KMS encryption instead has a different operating mechanism: AWS KMS helps create and control cryptographic keys used by AWS encryption features.

The interchangeable-service claim ignores that at-rest encryption protects persisted data with managed cryptographic controls. By contrast, AWS KMS helps create and control cryptographic keys used by AWS encryption features; the logistics review must therefore distinguish which function it actually needs to protect stored embeddings and documents against exposure of underlying storage media. Encrypted storage cannot prevent data leakage by an already-authorized application.

Choosing AWS KMS encryption to protect stored embeddings and documents against exposure of underlying storage media would confuse its intended workload with that of Encryption at rest. AWS KMS encryption is appropriate when a team needs to manage encryption keys protecting customer documents stored for an AI workflow; here the required behavior is that at-rest encryption protects persisted data with managed cryptographic controls.

The two roles have been exchanged in the description. For this industrial maintenance scenario, AgentCore Policy addresses the need to limit tool actions that an AI agent is allowed to perform; IAM role for workloads instead has a different operating mechanism: an IAM role provides temporary credentials to trusted AWS service workloads.

The interchangeable-service claim ignores that AgentCore Policy supports policy-based enforcement of agent actions. By contrast, an IAM role provides temporary credentials to trusted AWS service workloads; the industrial maintenance review must therefore distinguish which function it actually needs to limit tool actions that an AI agent is allowed to perform. Instructions in a prompt are not an enforceable substitute for tool authorization.

Choosing IAM role for workloads to limit tool actions that an AI agent is allowed to perform would confuse its intended workload with that of AgentCore Policy. IAM role for workloads is appropriate when a team needs to let a managed workload access AI-related AWS resources without embedded long-term keys; here the required behavior is that AgentCore Policy supports policy-based enforcement of agent actions.

Choosing Amazon CloudWatch monitoring to keep a retrieved page from issuing unauthorized instructions to an assistant would confuse its intended workload with that of Prompt injection defense. Amazon CloudWatch monitoring is appropriate when a team needs to observe model-backed application errors, latency and operating metrics; here the required behavior is that treat retrieved text as untrusted data and constrain tool actions with external checks.

The two roles have been exchanged in the description. For this university services scenario, Prompt injection defense addresses the need to keep a retrieved page from issuing unauthorized instructions to an assistant; Amazon CloudWatch monitoring instead has a different operating mechanism: CloudWatch collects metrics and operational logs for supported services and applications.

The interchangeable-service claim ignores that treat retrieved text as untrusted data and constrain tool actions with external checks. By contrast, CloudWatch collects metrics and operational logs for supported services and applications; the university services review must therefore distinguish which function it actually needs to keep a retrieved page from issuing unauthorized instructions to an assistant. Placing raw web content into a model prompt does not make it a trusted command.

The interchangeable-service claim ignores that grounding gives the model retrieved evidence relevant to the question. By contrast, A catalog records metadata, ownership and classification of data assets; the media publishing review must therefore distinguish which function it actually needs to reduce unsupported answers by providing approved source documents at inference. Retrieving a passage does not assure that every sentence in the response is correct.

The two roles have been exchanged in the description. For this media publishing scenario, RAG factual grounding addresses the need to reduce unsupported answers by providing approved source documents at inference; Data cataloging instead has a different operating mechanism: a catalog records metadata, ownership and classification of data assets.

Choosing Data cataloging to reduce unsupported answers by providing approved source documents at inference would confuse its intended workload with that of RAG factual grounding. Data cataloging is appropriate when a team needs to help data owners find the approved source and owner of each dataset; here the required behavior is that grounding gives the model retrieved evidence relevant to the question.

The interchangeable-service claim ignores that AWS KMS helps create and control cryptographic keys used by AWS encryption features. By contrast, TLS encrypts network communications and helps prevent interception in transit; the university services review must therefore distinguish which function it actually needs to manage encryption keys protecting customer documents stored for an AI workflow. Encryption does not determine whether a principal is authorized to retrieve decrypted data.

The two roles have been exchanged in the description. For this university services scenario, AWS KMS encryption addresses the need to manage encryption keys protecting customer documents stored for an AI workflow; Encryption in transit instead has a different operating mechanism: tLS encrypts network communications and helps prevent interception in transit.

Choosing Encryption in transit to manage encryption keys protecting customer documents stored for an AI workflow would confuse its intended workload with that of AWS KMS encryption. Encryption in transit is appropriate when a team needs to protect a customer's prompt while it travels between client and service; here the required behavior is that AWS KMS helps create and control cryptographic keys used by AWS encryption features.

The two roles have been exchanged in the description. For this banking operations scenario, Vulnerability management addresses the need to detect software flaws in infrastructure hosting an AI application; Amazon CloudWatch monitoring instead has a different operating mechanism: CloudWatch collects metrics and operational logs for supported services and applications.

The interchangeable-service claim ignores that vulnerability assessments and patching reduce exposure in software dependencies. By contrast, CloudWatch collects metrics and operational logs for supported services and applications; the banking operations review must therefore distinguish which function it actually needs to detect software flaws in infrastructure hosting an AI application. Model guardrails do not remediate exploitable container or host packages.

Choosing Amazon CloudWatch monitoring to detect software flaws in infrastructure hosting an AI application would confuse its intended workload with that of Vulnerability management. Amazon CloudWatch monitoring is appropriate when a team needs to observe model-backed application errors, latency and operating metrics; here the required behavior is that vulnerability assessments and patching reduce exposure in software dependencies.

The interchangeable-service claim ignores that an IAM role provides temporary credentials to trusted AWS service workloads. By contrast, AgentCore Identity helps agents use managed identity and credentials for authorized connections; the manufacturing quality review must therefore distinguish which function it actually needs to let a managed workload access AI-related AWS resources without embedded long-term keys. Hard-coded access keys can be exposed and require separate manual rotation.

Choosing Amazon Bedrock AgentCore Identity to let a managed workload access AI-related AWS resources without embedded long-term keys would confuse its intended workload with that of IAM role for workloads. Amazon Bedrock AgentCore Identity is appropriate when a team needs to manage agent authentication and delegated access to external resources; here the required behavior is that an IAM role provides temporary credentials to trusted AWS service workloads.

Choosing Resource-based policy to let a managed workload access AI-related AWS resources without embedded long-term keys would confuse its intended workload with that of IAM role for workloads. Resource-based policy is appropriate when a team needs to control which AWS principals can read a sensitive data bucket used for retrieval; here the required behavior is that an IAM role provides temporary credentials to trusted AWS service workloads.

The two roles have been exchanged in the description. For this manufacturing quality scenario, IAM role for workloads addresses the need to let a managed workload access AI-related AWS resources without embedded long-term keys; Amazon Bedrock AgentCore Identity instead has a different operating mechanism: AgentCore Identity helps agents use managed identity and credentials for authorized connections.