April 2026 Monarch Release Guide

Dashboards

New

An AI-generated summary widget for dashboards. A new AI natural language widget can be configured by a dashboard Author to generate a short text summary based on the data context the widget sits in. Dashboard viewers get an at-a-glance narrative — what the data is saying — alongside the charts and tables they already see.

Static Cards. A new Static Card component lets you place a fixed tile on a dashboard, with its own drilldown configuration, for interactivity that doesn't depend on a live data source.

Static URL drilldown targets. Any component can now be configured with a default static URL as a drilldown target — so navigation works on cards and components that have no live data behind them.

Improved

Drilldowns can open in a new tab or popup. Viewers exploring detail pages or external links no longer lose the dashboard they came from.

Dashboard tables. Hyperlinks change colour and underline on hover, drilldown cells highlight on hover so it's clear at a glance which cells are interactive, and the overall table style is more consistent with the rest of the dashboard.

Treeview behaviour. The treeview now mirrors the Browse area: drilling down is the default action when a node is opened, so you move through the hierarchy without having to click an "info" icon at each level. Clicking an already-active item keeps it active rather than deactivating it. The hierarchy diagram itself can be filtered by immediate parent, simplifying configuration of multi-level views.

Treeview sorting and labels. Items with number-prefixed labels (1. Discovery, 2. Plan, 11. Review) now sort in natural order rather than alphabetically, so 11. no longer appears between 1. and 2.. Long tooltips and titles handle truncation more gracefully than the previous fixed-width clipping.

Filter performance. Filter autocomplete is faster on large datasets, and the autocomplete overlay is now positioned correctly.

Linked content in components. Component data loading now also retrieves the datasources used by linked items, so links inside dashboard components resolve correctly on first render.

Model card list sorting is more predictable, with simplified sorting logic underneath.

Fixed

• Filters set to empty were excluding items where the field was null rather than empty string, so some items were missing from filtered views.
• Stale filters with no menu definition were being included in the filter count, so applied filter counts didn't match what was actually on screen.
• Charts using a count aggregation now fall back to the category field where appropriate, rather than requiring a separate numeric column to count rows.
• On aggregated chart drilldowns, the passing field value was being resolved incorrectly, so drilldowns sometimes carried the wrong context through to the destination.
• The treeview was double-firing drilldowns on items without children. Drilldowns now fire once.
• Widget backgrounds were rendering on hard-coded white rather than picking up the theme's surface colour.

Discovery Agent

Improved

A more familiar Discovery Agent UI. The Discovery Agent has had a round of UI standardisation, bringing its layout and components into line with the rest of the app — so jumping between the Discovery Agent and the rest of BusinessOptix feels less like switching products. A further round of refinements across the Agent has improved its overall behaviour and clarity.

Model as a Sheet

New

Column truncation with tooltip on overflow. Wide tables now truncate long values within the cell and reveal the full text on hover, so users no longer have to scroll sideways to read or work with the table.

Fixed

• Computed fields were throwing errors and not updating their value.
• A validation-driven memory leak has been resolved — clicking into a cell or selecting multiple rows in a large MaaS table no longer hangs the browser.
• Model Form was loading empty data when targeting a specific component via URL, instead of returning the requested component.

Forms

Fixed

• A locale-related issue could put forms with datetime fields into an infinite loop when the UI language was set to a non-UK locale; this has been resolved.
• Multi-language date validation now handles a wider range of locale formats reliably.
• Browser autocomplete is disabled on text and select inputs, preventing the browser's autofill from interfering with form values or obscuring real options.
• URL form fields were sometimes unopenable after being cleared because of a leftover focus trap; the focus trap has been removed and a duplicated update on edit has been simplified, so URL fields reopen cleanly after being emptied.
• Form dirty checking now compares against the original value to determine whether a node has actually changed, so the Save state correctly reflects whether anything has been edited.

Library navigation, search and output

New

Archived content in advanced search. Advanced search now includes an archived checkbox, with clearer search-result messaging when archived content is involved — so finding withdrawn or archived items no longer requires a separate, unrelated search.

Improved

File browser improvements. A wider round of polish to Browse navigation and folder-state handling.

Document drawer. Rounded corners on the document navigation drawer, plus a number of missing document table styles brought across so document outputs render closer to the live experience.

Hierarchy diagrams. Long unordered lists in hierarchy diagrams now use the correct padding.

Reading position with images. Documents containing images now adjust their scroll position when an image loads, so your reading position isn't lost when the page reflows.

User-settings avatar in the navigation drawer has been simplified for clarity.

Fixed

• The Browse sidebar was opening the wrong folder when navigating directly to a nested location, and the folder details icon could be missing on first load from a saved URL. Both have been resolved.
• Folder sort keys are now compared case-insensitively for consistent ordering.
• Renaming a model from advanced search results was erroring; it now works.
• Library-level facet overrides were not being applied correctly when settings were present, and facets set manually via URL parameters were not driving the right state. Both are now correct.
• The search dialog now closes automatically on route change, so the search box no longer pops back into focus on top of a model you've just opened.
• Opening a model from search could leave the loading state stuck or return 404s when the URL contained a path after the GUID.
• The Actions and Comments buttons in the app bar were not always present on the first load of a freshly opened model — a refresh is no longer needed. The Actions popup opens reliably again, and the show-actions logic that controls when buttons are visible has been corrected.
• The model type icon is displayed next to the title in the navigation bar for v7 models as well as v8, and we no longer issue duplicate API fetches resolving the icon.
• The page title is correctly populated on initial model load from a saved URL, including when only the model path is supplied. Pop-ups for sub-models no longer overwrite the main page title.
• Task card links now route to the correct destination.
• The v8 workflow alert in output for items awaiting review or approval now opens the v7 actions panel rather than doing nothing.
• The Comments modal no longer hides itself after tagging a user.
• Several remaining icons that weren't picking up theme colours have been corrected.