Amazon Athena now lets teams query service-managed tables of usage data, making it possible to build cost and activity analysis without maintaining a custom data pipeline. AWS also says its Agent Toolkit includes skills for querying those tables and surfacing usage patterns and optimisation recommendations.
AWS AI Watch analysis
What happened
The system tables record Athena usage data and make it available within minutes of a query completing. AWS says the read-only Apache Iceberg tables are stored on Amazon S3 Tables, with configurable retention and service-managed schema, partitioning, compaction and snapshot maintenance.
Teams can query the data directly in Athena. AWS gives cost-allocation dashboards, broken down by workgroup, user or table, as one example. Because the tables use Iceberg, they can also be queried with compatible tools including DuckDB, Redshift and EMR. The AWS announcement says the feature is available in regions where both Athena and S3 Tables are available, and can be enabled through the Athena console or API.
Why it matters
For platform teams, the practical change is access to query-usage data in a format they can analyse with SQL, without first building and operating their own extraction pipeline. That can make it easier to see where queries are running and what is driving data-scanning costs. The Iceberg format also means the data is not limited to Athena for analysis.
The AI connection is specific rather than decorative: AWS says its Agent Toolkit provides skills for querying the tables and producing usage patterns and optimisation recommendations. That gives teams an agent-assisted route to operational insight, though the announcement does not say how accurate or useful those recommendations are in practice.
Our read
This is a practical observability upgrade, especially for teams already working with Iceberg or trying to understand Athena costs across users and workgroups. The useful part is not that an agent can talk about usage; it is that the underlying data is queryable without a bespoke pipeline. Teams should check regional availability and retention settings before building workflows around it.
What to watch
- Whether the system tables expose enough detail for teams’ cost-allocation and audit needs.
- How AWS’s Agent Toolkit skills perform on real usage data, beyond the feature description.
- Whether teams can use the Iceberg tables effectively from their existing analytics tools.
Discussion spark: Would you let an AI agent recommend changes to your Athena workloads, or should it stick to surfacing usage data for people to interpret?
Sources and evidence
- Amazon Athena introduces system tables for observability, auditing, and usage analytics (9 November 2026, 16:00 UTC)
not affiliated with or endorsed by Amazon Web Services (AWS)