Skip to main content

​
Build Automations by Chatting with AI

You can now build entire automations just by describing what you want. A new AI Analyst panel sits right in the automation editor — open it from the toolbar or from the “Create with AI” option on an empty canvas, and describe the workflow you need (e.g. “email me when visits cross 500 in 5 minutes”). The assistant asks clarifying questions only when it truly needs them, builds out the trigger/action graph live on your canvas, and every change lands in your undo stack, so you’re always in control.

​
Conditional Branching for Automations

Automations can now make decisions. A new Branch (Switch) node lets you route a workflow down different paths based on conditions you define — for example, sending one email when a metric crosses a threshold and a different one otherwise. Each branch can check multiple conditions (joined with AND/OR) and you can add a fallback path for anything that doesn’t match.

​
Smarter, More Reliable Ask AI

  • Ask AI now understands relative dates correctly. Questions like “this month,” “this week,” or “YTD” are now resolved against today’s actual date in your timezone, instead of being guessed — so you get the right data range every time.
  • Error messages are cleaner. If something goes wrong while Ask AI is generating an answer, you’ll now see a plain, helpful message instead of raw internal error text.
  • A rare issue where some past search conversations failed to load has been fixed.

​
Dashboard & Chart Fixes

  • Fixed drilling into date-grouped charts. If you grouped a chart by month/week/day and then drilled into one of the bars, the drill-down previously came back empty. Drilling into a grouped bucket now correctly returns the underlying rows for that period.
  • Fixed several issues with decimal and number formatting in charts.
  • Fixed a recurring timestamp display issue across widgets.
  • Fixed a table pagination issue.

​
Database & Query Builder

  • Materialized views on Postgres now show their schema correctly, so dashboards built on top of materialized views work as expected instead of showing empty fields.
  • Improved reliability when reading from Postgres read replicas under heavy replication load — transient read conflicts are now retried automatically instead of failing the query.
  • MongoDB support in the AI Query Builder: the AI can now write and validate real MongoDB aggregation pipelines (not just simple queries) for your Mongo data sources, and Mongo queries containing $-operators no longer get corrupted.

​
Automation Reliability Improvements

  • PDF and PowerPoint reports generated by automations are now hosted on Supaboard’s own storage instead of a temporary third-party link, so exported files won’t expire.
  • Automated emails now show your workspace name in the header instead of always saying “Supaboard.”
  • Fixed cases where automation steps using upstream data (query results, AI answers) in an email body or AI insight step could silently send blank content.
  • Fixed automations timing out when an early step in a run was slow, even though the step you were testing was fine on its own.
Thank you for using Supaboard. We’re continuously improving the platform to help you explore, understand, and act on your data faster.
Last modified on July 21, 2026