Optimizing AI/BI Dashboard Performance: Best Practices for Speed and Scale
Overview
| Experience | In Person |
|---|---|
| Track | Analytics & BI |
| Industry | Enterprise Technology |
| Technologies | AI/BI, Databricks SQL |
| Skill Level | Intermediate |
| DOWNLOAD SESSION SLIDES | |
Slow queries, long load times, and inconsistent responsiveness don't just frustrate users — they limit adoption and erode trust in your data. In this session, we'll teach you how to apply the right design patterns and leverage the latest AI/BI Dashboard capabilities — including automatic materialization to deliver consistently fast, interactive performance at scale. You'll walk away knowing how to: - Design datasets and queries for fast, interactive dashboard performance - Leverage caching, query optimization, and warehouse configuration effectively - Use automatic materialization for the right queries to maximize performance gains - Identify and fix common performance pitfalls
Session Speakers
Daniel Streit
/Staff Software Engineer
Databricks
Eason Gao
/Senior Product Manager
Databricks
Full Summary
How to make AI/BI Dashboards load in seconds, not minutes
Slow dashboards erode trust and adoption. The session lays out a practical playbook for making AI/BI Dashboards feel instant, emphasizing that speed comes from stacking multiple techniques, not a single trick. The guidance focuses on precomputation and caching, smarter SQL and layout, and a new warehouse tuned for fast BI. Teams looking for broader context on how AI is transforming data analytics will find that the same principles of reducing waste and moving work earlier apply at every scale.
FAQ
No. When enabled on a metric view, automatic materialization reroutes queries to the materialized layer transparently. Existing queries and governance remain unchanged, and permissions are inherited from the metric view.
Pick unaggregated when joins or filters are expensive and you want to retain row‑level detail. Choose aggregated when dashboards repeatedly group by the same dimensions and measures, so those rollups can be served directly without re‑aggregating at query time.
Set a schedule using the most common default parameters to warm the shared server cache for initial loads. If you author datasets in SQL, check for the Optimize button in Genie and accept suggested improvements that meet the projected speedup.
Those limits reflect what can be fetched and queried locally while preserving a snappy experience. When a dataset fits, filtering happens in the browser and feels instant. Larger datasets rely on server execution, where materialization and SQL optimization matter more.
It is in private preview. Contact your Databricks account team to request access and evaluate it on your canonical dashboards as the team validates real customer query patterns.