Contents
The first half of 2026 went on two things: making the numbers in SwiftCase easier to get at, and making the record of what happened easier to search. Here's what shipped.
A statistics dashboard you configure yourself
Management information used to mean the reports we'd built for you. Now it means the ones you build yourself.
The dashboard has been generalised. Pick what you want to count, pick the field to count it by, add clauses until the query describes the question you actually have, and save it. Results come back as charts, with date filtering on top and CSV export underneath for when the answer has to leave SwiftCase and go into a board pack.
The charts are the visible bit. The change that matters is that a question you asked once becomes a number you can watch, and nobody had to raise a ticket for it.
Logs you can actually search
Two log improvements landed together and complement each other.
System log search turns the audit record into something you can interrogate. When a client asks what happened on a case back in March, you run a search. It used to be an afternoon.
The task log view was reworked, with CSV export alongside it. In a complaint or an audit, nobody is really asking what the current status is. They're asking how it got there, and the task log is where that lives. Now it exports cleanly for anyone who needs it outside the platform.
Error logs got the same treatment: a rebuilt index and proper reporting, so an operational problem surfaces as something readable instead of a wall of stack traces.
Guided walkthroughs
New starters learn a configured workflow by being shown it, usually by the person who can least afford the hour. SwiftCase now runs guided walkthroughs directly in the interface. The tour points at the real screen, on your real configuration, so nobody is reading a screenshot that went stale two releases ago.
The same mechanism drives training tours and interface tests, so a walkthrough breaks loudly the moment the page it describes changes. Documentation that can fail a build stays true.
Email, comprehensively
Email had the busiest half of any part of the platform.
Drafts are now managed properly, including recreating a draft from its template when the underlying template has moved on. Email domains have their own admin section. Templates gained a fuller management interface. Email threads are retrievable through the API, so an integration can pull the whole conversation on a case rather than one message at a time. Attachment handling was reworked, along with the interface around it.
There's also a clearer error when a task tries to send from a user with no email provider configured. Sounds trivial. It was one of the most common support questions we had.
Automation rules that read the answers
Global automation rules can now compare against dynamic task answers as well as static values. A rule can say "when this answer exceeds that answer" instead of "when this answer exceeds 500", which kills off a whole category of near-identical rules that existed only because the threshold varied.
Scheduled events gained an admin interface and, importantly, duplicate-scheduling prevention. A rule that fires twice does more damage than one that doesn't fire at all, because the second firing stays invisible until a client asks why they got two of everything.
Case relationships became additive
A smaller change with a disproportionate effect. Case relationships are now additive. Add a new relationship to the database, say a valuer, and it's immediately available in templates as a variable. No code change, no release.
Anyone who's ever waited on a deployment to add one field to one letter will understand why it's in the list.
Time logs
Time logs can now be edited and deleted, and filtered by user state, which matters the moment someone leaves and their historical time still has to be reportable. Bulk user creation gained a direct insert path for the volumes where the ordinary route was too slow, and parent user now comes through on CSV export.
Under the surface
Two pieces of groundwork that will show up in everything that follows.
A shared component library now underpins the interface: modals, toasts, tooltips, tabs, date rendering and a declarative form renderer, all in one place instead of reimplemented per page. On top of it, most of the interface moved to Turbo Frames, so notes, fees, documents and flags update in place without reloading the page.
The automation rules editor was rebuilt on that library, and every rule now carries a plain-English summary of what it actually does. Reading a rule and understanding it ought to be the same activity.
Further reading:
- Changelog: what shipped, month by month
- Documentation: how the platform is organised, and how to configure it
- Analytics and reporting: the reporting side of the platform
- Workflows: building and changing the processes your work runs through