Amazon SageMaker Data Agent now available for IAM Identity Center domains
Data Agent now works in enterprise IAM Identity Center domains—letting analysts generate SQL/Python and debug code using plain English.
View original announcement →Visual Summary
graph TD
A{{SageMaker Data Agent}}:::announced
B((Data Analyst)):::external
C([Natural Language to Code]):::feature
D([Fix with AI Debugging]):::feature
E(SageMaker Notebooks):::compute
F(Query Editor):::compute
G(Amazon Athena):::storage
H(Amazon Redshift):::storage
I(Amazon S3):::storage
J(AWS Glue Data Catalog):::storage
K([IAM Identity Center]):::feature
B ==>|"plain English"| A
K -->|"authenticates"| A
A -->|"generates code"| C
C -->|"Python"| E
C -->|"SQL"| F
A -.->|"analyzes errors"| D
E -->|"queries"| G
F -->|"queries"| H
E -->|"reads"| I
F -->|"discovers"| J
classDef announced fill:#ff9900,stroke:#ec7211,color:#fff,font-weight:bold
classDef compute fill:#e3f2fd,stroke:#1565c0,color:#1565c0
classDef storage fill:#e8f5e9,stroke:#2e7d32,color:#2e7d32
classDef feature fill:#fff3e0,stroke:#e65100,color:#e65100
classDef external fill:#f5f5f5,stroke:#616161,color:#616161
What's New
Amazon SageMaker Data Agent is now generally available for SageMaker Unified Studio domains configured with AWS IAM Identity Center (IDC), expanding its reach to enterprise identity-managed environments. The agent enables data analysts and engineers to describe analytical goals in plain English and receive working Python or SQL code tailored to connected data sources such as Amazon Athena, Amazon Redshift, Amazon S3, and AWS Glue Data Catalog. It operates across both SageMaker notebook and Query Editor environments, and includes a conversational context engine and an AI-powered "Fix with AI" debugging feature.
How It Works
- Natural Language to Code: Users describe their analysis goals in plain English; the agent generates executable Python or SQL code targeting the connected data sources (Athena, Redshift, S3, Glue Data Catalog) without requiring manual query authoring.
- Step-by-Step Planning: Before generating code, the agent proposes an incremental plan for the analysis, giving users the opportunity to review and adjust the approach before execution.
- Conversational Context: The agent maintains state across notebook cells, selected tables, and query history, enabling multi-turn interactions where follow-up requests build on prior context rather than starting fresh.
- Dual Environment Support: Data Agent is accessible from both SageMaker notebook environments (for Python/DataFrame workflows) and the Query Editor (for SQL-centric analytics), covering the full analytics development lifecycle.
- Fix with AI Debugging: When code execution fails, the "Fix with AI" feature analyzes the error output and suggests targeted corrections, reducing the iteration cycle for debugging complex queries or transformations.
- IAM Identity Center Integration: The agent is now available in Unified Studio domains authenticated via IAM Identity Center, enabling enterprise SSO-managed users to access the capability within their existing identity governance framework.
Why It's Important
- Reduces Skill Barrier: Data analysts without deep SQL or Python expertise can now perform complex aggregations, joins, and DataFrame transformations through natural language, democratizing access to advanced analytics.
- Accelerates Development Cycles: Automated code generation and AI-assisted debugging significantly reduce the time from analytical question to working code, compressing what could take hours into minutes.
- Enterprise Identity Readiness: Support for IAM Identity Center domains means organizations using centralized SSO and permission sets can now adopt Data Agent without compromising their identity governance or access control policies.
- Contextual Continuity: Maintaining conversational context across cells and query history eliminates the repetitive re-specification of data context, making iterative analysis more natural and efficient.
- Unified Analytics Surface: By spanning notebooks and Query Editor, the agent removes the friction of context-switching between tools, keeping analysts in a single environment regardless of whether they prefer Python or SQL.
How It's Different
- IDC Domain Support: Unlike the initial release, this expansion explicitly supports IAM Identity Center-configured domains, which is the standard enterprise deployment model for SageMaker Unified Studio—making this the first time the feature is accessible to most enterprise customers.
- Multi-Source Awareness: The agent is natively aware of multiple connected data sources simultaneously (Athena, Redshift, S3, Glue), whereas generic AI coding assistants lack the catalog and schema context needed to generate accurate, runnable queries.
- Plan-Before-Generate Paradigm: The agent proposes a step-by-step plan prior to code generation, a deliberate design choice that differentiates it from tools that immediately emit code, giving users more control and transparency.
- Integrated Debugging Loop: The "Fix with AI" feature is tightly integrated with the execution environment, meaning it can read actual runtime errors in context—unlike standalone LLM chat interfaces where users must manually copy-paste error messages.
- Notebook Cell Context Awareness: Unlike standalone SQL or code assistants, Data Agent tracks the state of prior notebook cells and selected tables, enabling coherent multi-step analytical workflows rather than isolated one-shot completions.
When to Prefer It
- Enterprise SSO Environments: Use Data Agent when your SageMaker Unified Studio domain is configured with IAM Identity Center, as this release specifically unlocks the feature for that deployment topology.
- Ad Hoc Analytics on Complex Schemas: Prefer Data Agent when analysts need to query large, multi-table schemas in Athena or Redshift where manually constructing joins and aggregations is error-prone and time-consuming.
- Iterative Exploratory Analysis: Ideal when the analytical question evolves over time (e.g., drilling into quarterly revenue trends), as the conversational context engine handles follow-up refinements naturally.
- Mixed Python and SQL Workflows: Use it when a project requires both DataFrame transformations in notebooks and SQL queries in the Query Editor, benefiting from a single agent interface across both surfaces.
- Rapid Prototyping and Debugging: Prefer Data Agent when developers need to quickly prototype analytics pipelines and want AI-assisted error correction to reduce debugging overhead.
- Enabling Non-Expert Users: Adopt Data Agent when business analysts or domain experts need to self-serve on data without relying on data engineering teams to write bespoke queries.
Availability
- GA Status: Generally available as of May 13, 2026; this is not a preview release.
- Supported Regions: Available in all commercial AWS Regions where Amazon SageMaker Unified Studio is supported; no specific regional exclusions were noted beyond non-commercial partitions.
- Domain Requirement: Requires a SageMaker Unified Studio domain configured with AWS IAM Identity Center; domains using other authentication configurations may not yet be supported.
- Access Point: Accessed via the Data Agent panel within a project's notebook or Query Editor in SageMaker Unified Studio—no separate service enrollment required.
- Pricing: No separate pricing was announced; costs are expected to follow SageMaker Unified Studio and underlying service consumption pricing (Athena queries, Redshift usage, etc.); verify current pricing on the AWS pricing page.
- Data Source Support: Confirmed integrations include Amazon Athena, Amazon Redshift, Amazon S3, and AWS Glue Data Catalog; other data sources are not mentioned in the announcement.
