Amazon Q is a family of generative AI assistants rather than a single endpoint with one configuration model. For engineering teams, the main distinction is between Q Developer, which supports software development and AWS workflows, and Q Business, which provides an enterprise assistant grounded in organizational information. Select the offering according to the task, then evaluate its supported Regions, identity requirements, administrative controls, and pricing independently.
Q Developer is available through supported development environments and AWS interfaces. Engineers can use it to explain unfamiliar code, generate implementation suggestions, write tests, investigate errors, and obtain guidance about AWS resources. Certain capabilities can perform multi-step development tasks or assist with supported code transformations. Availability depends on the interface, subscription, programming language, and task. Treat generated code and infrastructure changes as proposed work: review the diff, run automated tests, and apply the same security and release controls used for human-authored changes.
Q Business is configured around an application, an identity setup, and access to enterprise content. A typical deployment connects repositories or object storage, configures ingestion and synchronization, and exposes a conversational interface to authorized employees. Responses can reference retrieved documents so users can inspect the material supporting an answer. Content freshness depends on successful ingestion and synchronization; changing a source document does not necessarily make that change immediately available in responses.
Authorization requires careful design. For Q Business, configure user and group identities and preserve supported document-level access controls when ingesting content. Validate access with representative users, particularly when connectors map identities across different systems. For Q Developer, distinguish the permissions needed to access the assistant from the AWS permissions needed to inspect resources or execute operations. An answer about an AWS service is not evidence that the caller is authorized to change that service.
Security and governance settings differ across offerings and subscription tiers. Review data handling, encryption, retention, supported administrative policies, and any settings controlling the use of content for service improvement before enabling sensitive workloads. Do not place credentials or secrets in prompts. Where actions or external integrations are enabled, restrict execution roles and require appropriate confirmation or approval for consequential changes.
Operate each deployment with measurable acceptance criteria. For Q Business, test retrieval quality, document permissions, synchronization failures, and citation relevance using a representative question set. For Q Developer, evaluate suggestion quality, test coverage, security findings, and developer workflow fit. Budget separately for subscriptions, indexed capacity where applicable, and supporting AWS resources. Generated answers can be incomplete or incorrect, so retain human review for production decisions.
AWS CloudFormation
Amazon S3