September was a busy month product-wise for Arize AX, with updates to make AI agent engineering faster and easier. From session and trace evals to Alyx’s new synthetic data generation skill, there is a lot to catch up on and try out.
Session and Trace Evals
Session- and trace-level evaluations are available across all Arize tiers. This unlocks the ability to evaluate at different levels (span, trace, and session), which is crucial for building and monitoring agents.
Alyx: Synthetic Data Generation Skill
You can now generate synthetic datasets directly in Playground using Alyx, Arize AX’s assistant with agent mode. This makes it easier to quickly create sample data for testing and experimentation without needing external tools.
Annotations Updates
Annotations Config: Optimization Direction
Added support for defining optimization direction on categorical and continuous configs. The optimization_direction field now lives within their respective config columns, improving logical grouping.
Annotations Autosave
Annotations on the Experiment Compare page now autosave immediately after selection, removing the need to manually save them. Experiment annotations are also displayed on the experiment compare page.
Experiment Annotations on Experiment Compare Page
Annotations are now visible directly on the Experiment Compare page, making it easier to review context and insights alongside your experiment results. This streamlines analysis by keeping annotations and metrics in one place.
Dashboard Improvements
Dashboards have been reworked with a richer experience: clickable legends and new widget creation forms for line charts, bar charts, experiments, monitors, and statistics.
Dashboards also now support per-widget time range selection and improved legend display options, giving users more granular control and clearer data visualization. These updates make it easier to tailor each dashboard view to your specific analysis needs.
Users can also now position widgets freely, including adding gaps for layout flexibility.
Other Updates
Improved Sessions View
Text in Session bubbles now wraps automatically, ensuring all content stays visible for a smooth sessions experience.
Large Dataset Runs in Playground
The prompt playground now supports running on an unlimited number of rows, with a new “Load 100 more” button to make adding records easier. This is the first iteration of scalable runs, and we’ll continue improving the usability in upcoming updates.
Search In Settings
All settings tables now include search bars, making it easier to quickly find what you need.
Want us to show you any of these features? We’d love to schedule some time!