Enterprise Reporting and Embedded Analytics with AI/BI Dashboards

The latest updates to AI/BI Dashboards, scheduling, embedding, and Genie focus on dependable delivery, strong data permissions, and simpler self-service, reflecting patterns seen across modern business intelligence and analytics work. Teams looking to understand the broader shift in how AI is reshaping this space will find useful context in how AI is transforming data analytics.

The conversation highlights three distribution surfaces that work together: scheduled reporting for predictable delivery, embedded dashboards that put analytics inside applications, and Genie for conversational analysis and recurring reports users create themselves.

Governed scheduling and delivery options

Scheduled reporting remains central for teams that rely on daily and weekly briefings. Running schedules before business hours warms the query result cache, so dashboards load quickly when users arrive. Every schedule honors Unity Catalog permissions, including individual data permissions and row-level security, so each recipient sees only their entitled slice in the snapshot and attachments.

Delivery reaches users in email, Slack, and Microsoft Teams. Snapshots include PNG and PDF by default, and authors can attach specific widgets as CSV, TSV, or Excel for downstream analysis. Schedules can send only selected dashboard pages, keeping reports focused. Filters can be applied at schedule time to snapshot the exact view intended.

For delivery keyed to data freshness, Lakeflow Jobs can trigger a dashboard task when Unity Catalog tables update. Jobs can run conditionally based on upstream task output, and push parameters at the task or job level so refreshes and notifications are tailored. Standard job notifications provide visibility into failures and completions across the workflow, including the dashboard step.

Embedding for internal tools and customer portals

Embedding places AI/BI Dashboards directly inside the applications where teams already work. Internal embedding uses a simple iframe to drop a fully interactive dashboard, including the attached Genie agent, into tools such as Confluence. Unity Catalog governance flows through, so users see only what they are permitted to see.

External embedding, now generally available, extends interactive dashboards to end customers who do not have a Databricks account. An application you control authenticates with a service principal via OAuth. An AI/BI Dashboards external value parameter passes identity-aware context from your app into the dashboard query, providing per-viewer data isolation without provisioning each user in Databricks.

Genie for conversational and recurring reporting

Genie is a simplified home for data consumers. Users can browse certified dashboards, agents, and apps, and ask questions in natural language. Genie searches workspace knowledge, selects relevant data, runs SQL, and returns clean analyses with metrics and takeaways.

A key shift is automated recurring reports. In chat, a user can turn an analysis into a scheduled task. Tasks can also be created manually. Each task includes editable instructions, a cadence, optional MCP connectors for pulling external data or taking actions, and a run history for inspecting executions. Genie uses the latest data each run and, based on the prompt, decides whether to notify the user. Reports arrive as well-formatted emails, turning a question into a monthly or weekly briefing without filing a ticket.

Turning analytics into reliable, personalized delivery

The updates close gaps between producing insights and getting them to the right people. Schedules deliver governed snapshots in the channels and file formats teams use, embedding carries secure interactivity into apps for employees and customers, and Genie lets business users create recurring, tailored reports from a conversation. Rapid customer adoption, such as Rivian's multi-domain migration that helped shape dozens of features, signals how quickly teams can consolidate reporting on this stack.