Business Readiness and AI Transformation

Question 65

A healthcare provider is preparing a diagnostic-imaging initiative. The technology team has configured AWS Organizations, prepared an Amazon SageMaker pilot, and proposed Amazon QuickSight dashboards. However, clinical leaders have not agreed on risk tolerance, clinical ownership, success criteria, or funding authority. What should the organization do before approving the pilot?

A. Purchase additional GPU capacity so the technology team can begin training immediately.

B. Create a centralized data lake before determining who owns clinical outcomes and funding decisions.

C. Select a pretrained medical service and defer business ownership decisions until after deployment.

D. Appoint an accountable executive sponsor and align clinical and business leaders on outcomes, risk tolerance, ownership, and funding authority.

Question 66

An enterprise reports several isolated Amazon SageMaker Studio experiments. Teams use different tools, there is no standardized model registry, deployment processes vary by team, and no shared success metrics exist. Which two observations best indicate that the enterprise is in an experimentation or emerging AI-maturity stage? Select two.

A. The enterprise uses centralized model lineage, automated CI/CD, and approved deployment templates.

B. Business units run ad hoc pilots with team-specific tooling and inconsistent evaluation practices.

C. A central registry requires approval for every version and provides standardized rollback procedures.

D. Teams have limited production governance and no repeatable intake process for AI initiatives.

E. Enterprise monitoring automatically tracks drift, quality, cost, and business outcomes for every model.

Question 67

A manufacturing company has a successful Amazon SageMaker proof of concept with a strong F1-score. The company has no MLOps engineers, no production support owner, no incident runbooks, and no approved deployment process. Which capability gap is the primary barrier to production?

A. The company needs more training data before it can assess the model.

B. The company lacks operational and people capabilities for deployment, monitoring, support, and incident response.

C. The company should replace the model with an Amazon Managed AI Service.

D. The company should increase endpoint instance size to improve production readiness.

Question 68

A logistics company wants to reduce delivery delays. Shipment data is reliable, but the company lacks real-time event ingestion, data science skills, model-risk review, and integration with dispatch workflows. Which three investments should it prioritize? Select three.

A. Assign operational ownership so dispatch teams can incorporate forecasts into their processes.

B. Rebuild the shipment data catalog even though shipment data is already reliable.

C. Build Amazon SageMaker ...