//
Search across all documentation pages
6 pages in this section.
A mental model for AWS's AI offerings: Amazon Bedrock for managed foundation models, Amazon SageMaker AI for building your own, and purpose-built services for ready-made vision, speech, language, and document tasks.
A starting tour of AWS's AI landscape: the three layers, which service fits a first project, and how to make your first call to Bedrock, Rekognition, Comprehend, and Polly from both SDKs.
When a task-specific AWS AI service (Comprehend, Rekognition, Textract) beats a general foundation model on Bedrock, and when the flexibility of a foundation model wins - with the trade-offs on cost, accuracy, and effort.
Why generative AI has become the fastest-growing region of AWS's AI surface in 2026, how Amazon Bedrock turned foundation models into ordinary API calls, and what that means for how you build - kept qualitative, not a market forecast.
A decision checklist for picking the right AWS AI service per use case: start at the highest layer that fits, prefer purpose-built for fixed tasks, use Bedrock for open-ended work, control token and endpoint cost, and secure model access.