Introduction
Since saying the Normal Availability of AI/BI Genie and introducing Databricks One at our most up-to-date Knowledge + AI Summit, we’ve continued to enhance how organizations analyze and act on information. This month’s roundup highlights new capabilities in AI/BI Dashboards and Genie — from deeper semantic tooling to simpler customization and higher consumer expertise.
These updates are a part of our mission to make enterprise intelligence extra correct, versatile, and accessible. For context, Databricks AI/BI is a built-in suite of BI capabilities in Databricks SQL that lets customers ask questions in pure language, construct dashboards, and share insights — all powered by the Databricks Knowledge Intelligence Platform.
What’s New?
As standard, let’s begin by a number of the main new headlines for each AI/BI Dashboards and Genie.
Dashboard Themes
Theme customization—one of the steadily requested options from clients—is now accessible in AI/BI Dashboards. Authors can now configure coloration palettes and font kinds to higher align dashboards with their group’s branding requirements. These settings could be adjusted from the dashboard’s settings panel, the place customers can select a preset theme or outline a customized one—together with colours for textual content, background, widgets, and chart collection, in addition to title alignment and font kinds. Theme adjustments apply persistently throughout your entire dashboard and are preserved throughout export and import, enabling scalable, brand-aligned reporting throughout groups.
International Dashboard Filters
Dashboard authors can now outline International Filters that apply throughout all pages of a dashboard. Accessed by clicking the filter icon within the dashboard toolbar, International Filters open a configuration panel the place you possibly can add interactive filter widgets scoped throughout your entire dashboard. Select from field-based filters (e.g., date pickers, categorical dropdowns) or parameter-driven filters, and specify default values to regulate conduct on preliminary dashboard masses. This characteristic dramatically simplifies multi-page dashboard design—enabling constant, intuitive navigation and decreasing redundant setup—whereas giving customers a centralized filter expertise.
Dashboard Drill By way of
Drill By way of gives customers a strong new option to discover dashboards by making use of alternatives from one web page as filters on one other. With this characteristic, viewers can right-click on an information level in a supply visualization—equivalent to a bar, pie chart and navigate to a goal web page the place their choice is utilized as a neighborhood filter. For instance, deciding on “United States” in a sales-by-country bar chart on a abstract web page can robotically filter the element web page to indicate solely U.S. information. Drill By way of unlocks new workflows for investigative evaluation, serving to customers transfer seamlessly from high-level KPIs to deeper diagnostic insights without having to reset filters or manually repeat alternatives.
Dashboards File Add
Now you can add Excel, CSV, or TSV information straight into Unity Catalog from the Dashboards information tab. To outline a dataset from a file, after deciding on a dashboard draft, navigate to Knowledge, select Add file, and specify the goal Catalog, Schema, and Desk title. You select whether or not to create a brand new desk or overwrite an current one—simply guarantee you’ve gotten MANAGE privileges on the goal schema. As soon as uploaded, the file turns into a first-class dataset: question it utilizing SQL, visualize it instantly, or mix it with different Unity Catalog information—all with out leaving the dashboard interface. This empowers customers to discover advert hoc or exterior information seamlessly, decreasing friction within the analytics workflow.
Forecasting for Dashboard Line Charts
Now accessible in Public Preview, AI/BI Dashboards now leverage the ai_forecast operate to let customers generate forecasts straight from line charts, remodeling historic developments information into forward-looking insights. To make use of it, choose any line chart with a temporal x‑axis and a numeric y‑axis, then click on the “+ Forecast” button within the visualization editor and select “Add AI Forecast”. This creates a brand new chart that seamlessly overlays a forecasted pattern line. The underlying ai_forecast operate calculates projections based mostly in your chosen information vary and aggregation logic. This characteristic makes it straightforward to anticipate future patterns—whether or not that’s gross sales efficiency, internet site visitors, or operational metrics—and incorporate proactive perception into your dashboards.
Enhanced Genie Information Retailer
AI/BI Genie’s data retailer is a foundational development for enhancing the standard, relevance, and explainability of solutions returned by Genie. Obtainable inside every Genie house, the data retailer provides analysts a set of instruments to fine-tune how Genie understands information—enabling extra reliable pure language interactions with out having to switch upstream Unity Catalog belongings to start out curating the house.
Well-liked queries at the moment are robotically surfaced when new tables are added, giving enterprise customers prompt Instance SQL based mostly on steadily run queries that the writer has entry to.
Authors may curate and edit dataset metadata for his or her Genie house with a number of new capabilities:
- Native desk and column descriptions enable authors to rapidly iterate on desk metadata inside a Genie house, without having to overwrite the descriptions in Unity Catalog. That is significantly useful when an writer lacks permission to switch Unity Catalog belongings, needs to check totally different descriptions earlier than making a modification, or needs descriptions that enhance Genie accuracy however don’t belong as a central definition.
- Column synonyms make it simpler for Genie to map various enterprise phrases to the right technical fields.
- Column present/conceal lets authors declutter the semantic floor by excluding duplicate, irrelevant or overly technical fields, with out requiring new tables or views.
- Instance values assist Genie higher perceive how your information is formatted throughout all totally different column sorts (strings, numerics, dates, and many others).
- Worth dictionaries assist Genie map consumer prompts to distinctive string values in your information (e.g. mapping a consumer immediate for “UK” to “nation = United Kingdom”). Genie house authors can now add worth dictionaries for as much as 60 columns and 1,024 distinctive values per column.
- Be a part of relationships can now be declared for any dataset to specify joinable datasets, SQL JOIN expressions, and be part of sorts for use. This helps authors who lack permissions to set Major/International Keys in Unity Catalog or are working with views that don’t assist Major/International Keys. Current Major/International Keys from Unity Catalog will robotically sync to cut back threat of duplication/battle.
Collectively, these new options give authors considerably extra management over Genie’s conduct and allow extra correct, context-aware responses—marking an essential step towards delivering ruled, dependable, and business-aligned conversational analytics. These native variations of desk metadata let house authors rapidly iterate and outline context helpful for Genie without having to replace Unity Catalog globally. For extra info on Genie’s data retailer, please take a look at the product documentation.
Different notable options
We’re consistently working to increase and enhance AI/BI Dashboards and Genie with new options and capabilities. Our ongoing efforts spotlight our dedication to creating AI/BI extra user-friendly, sturdy, and impactful in supporting your group’s data-driven decision-making. Some noteworthy updates are included beneath; for a whole listing of options and fixes, please check with the AI/BI product launch notes.
AI/BI Dashboards
- Choropleth Maps. Choropleth maps enable information to be visualized geographically by coloring areas based mostly on aggregated values. Ideally suited for displaying patterns like income by state or incidents by nation, this new map sort helps U.S. and worldwide geographies and provides a spatial layer to decision-making in dashboards.
- Parameterized Dashboard Schedules. Authors can now outline named dashboard schedules with customized filter parameters, enabling focused report supply with out duplicating dashboards. That is splendid for groups distributing reviews to totally different areas, enterprise items, or buyer segments — all whereas sustaining a single supply of fact and decreasing upkeep overhead. For extra info on creating and sustaining schedules, please check with the product documentation.
- Schedule With out Embedded Credentials (Run as Viewer). Scheduled dashboard deliveries can now be executed utilizing the recipient’s viewer permissions as an alternative of embedding the writer’s credentials. This enhances safety and entry governance, making certain that every consumer sees solely the info they’re entitled to—and aligns supply conduct with enterprise compliance insurance policies.
- Outline Customized Calculations Over a Vary: AGGREGATE OVER measures at the moment are supported in customized calculations, empowering analysts to create rankings, operating totals, or shifting averages. This removes the necessity to outline window features throughout the SQL question straight, enabling filtering and aggregations on the identical base dataset throughout a number of widgets.
AI/BI Genie
- Editable Outcomes Filters. Genie customers can now refine Genie solutions by enhancing filters straight throughout the outcomes panel. For instance, after receiving a solution filtered by a selected area or product, customers can rapidly modify the filter values and regenerate outcomes — serving to customers rapidly appropriate queries and accelerating time to perception.
- Level Maps Now Supported. Genie now helps level maps in visible responses, enabling customers to view geographic information plotted as particular person factors based mostly on coordinates. This enhancement brings location intelligence to conversational analytics and helps customers visualize patterns like retailer places, service requests, or supply routes.
- Comply with-up Questions for Textual content Responses. Genie now suggests follow-up questions for text-based responses that don’t embody generated SQL or consequence units. Beforehand, follow-up prompts appeared solely after responses with information output. Now, even responses like “Clarify the dataset” embody prompt follow-up questions to assist customers uncover new questions they’ll ask and deepen their exploration inside a Genie house.
- Genie API Listing and Delete Operations. New API endpoints let builders programmatically listing and delete Genie areas and conversations inside an area— serving to organizations automate governance, handle content material at scale, and combine Genie into current workflows and tooling. These APIs are an essential basis for groups rolling out Genie throughout massive enterprises.
What’s Subsequent?
We’re persevering with to construct towards a extra unified, clever, and accessible analytics expertise throughout Databricks AI/BI. Right here’s a fast have a look at what’s coming:
- Enterprise semantics throughout Genie and Dashboards — Already in Public Preview for Dashboards, Metric Views will quickly energy Genie as nicely, bringing trusted KPIs into pure language workflows and shutting the loop between visible and conversational analytics.
- APIs for embedded Genie experiences — New APIs will allow you to export/import areas and accumulate reply suggestions, unlocking scalable deployment and tighter integration into your personal apps.
- Smarter Dashboards and simpler sharing — Quickly you’ll see a redesigned counter visualization, a brand new Apply button for filters, and assist for CSV attachments in e mail subscriptions — all designed to spice up usability and attain throughout your group.
We’ll be diving deeper into AI/BI semantics in upcoming blogs and assets — so keep tuned!
Be taught Extra
For those who’re trying to expertise the most recent AI/BI capabilities firsthand, the choices beneath are designed to information you thru.
- Free Trial: Get hands-on expertise by signing up for a free trial.
- Documentation: Dive deeper into the small print with our documentation.
- Webpage: Go to our webpage to study extra.
- Demos: Watch our demo movies, take product excursions and get hands-on tutorials to see these AI/BI in motion.
- Coaching: Get began with free product coaching via Databricks Academy
- eBook: Obtain the Enterprise Intelligence meets AI eBook