Somebody on your platform team built a workaround for this. I'd bet on it.
Maybe it was a UI hack that faked a pinned column so fulfillers could scroll right without losing track of which record they were on. Maybe it was a custom widget that summed a column because someone in finance asked a question the list couldn't answer. Maybe it was a related-list workaround so an agent could see a catalog request's items, tasks, and approvals without opening four tabs. Whatever it was, it worked, it was never documented, and it is still sitting in your instance today.
The Australia release just made a chunk of that redundant.
ServiceNow's Developer Passport session for Australia walked through the Forms and List enhancements landing in workspace. It's a longer list than the headline AI features get credit for — ten of them — and almost all of it is the unglamorous plumbing practitioners have been faking for years.
Here's what shipped, why two of these matter more than they look, and what I'd go delete.
What actually shipped
On forms:
- Field background color highlighting
- A rich text editor for Work Notes and Additional Comments, with a toggle icon to enable or disable it
- Format Painter and Email Compose updates
In the activity stream:
- Filter activity by "Posted By"
- Entries authored by AI are flagged as AI
In lists:
- Freeze and pin columns
- Inline visualizations — right-click a column and the chart renders on the same page
- List calculations — sum, average and more, directly in workspace
- Hierarchical related lists for catalog requests, items, tasks and approvals
- A stage details icon for workflow fields in list view
The detail sitting underneath all of it matters more than any single item: these are platform-level, delivered through the List Bundle that most out-of-box workspaces already use. That's the difference between a feature you adopt and a feature you inherit. Nobody has to build a business case for pinned columns. They just show up.
The AI attribution flag is the sleeper
Nine of these are quality-of-life. One is governance, and it's the one nobody will put in a slide.
Entries authored by AI now get flagged as AI in the activity stream, and you can filter activity by who posted it.
Read that again with a Now Assist deployment in mind. Every organization running agentic features has a question it has mostly been answering with a shrug: when an agent writes a work note, how does the next human know a human didn't write it? Six months from now, when someone is reconstructing why a P1 was closed the way it was, "the work notes said the root cause was identified" means something very different depending on whether a person or a model typed it.
You cannot audit what you cannot attribute. Attribution is not a nice-to-have on top of agentic automation — it is the thing that makes the rest of it reviewable. Every governance conversation I've had about AI on the platform eventually lands on some version of "but how would we know?" This is a piece of the answer, shipped in the platform rather than bolted on, and it arrived quietly in a list of UI improvements.
If you're running Now Assist in production, this is the item on the list to go look at first.
Freeze columns is not a cosmetic feature
The other one worth defending: pinned columns is filed under convenience, and it isn't.
Watch someone work a wide list. They scroll right to find State or Assigned to, and the identifier — the thing that tells them which record this is — slides off the left edge. Now they're counting rows. Row four. Probably row four.
That's not an annoyance, it's an error rate. It's how the wrong incident gets updated, and it's a class of mistake that never shows up in a defect report because it doesn't look like a bug. It looks like someone being careless. The platform stopped making people count rows, and the honest read is that this should have shipped years ago — which is exactly why the community reaction has been "finally" rather than "nice."
List calculations quietly attack the export habit
Sum and average, directly in the list, is a small feature aimed at a large problem.
Right now, when someone needs a total, one of two things happens: they export to Excel, or they file a request for a report. The first creates a number that lives outside the platform with no lineage and no refresh — it's true on Tuesday and wrong by Thursday, and someone will paste it into a deck anyway. The second creates a queue.
Both are the same failure: the platform holds the data and can't answer the question, so the answer moves somewhere the platform can't govern. Inline calculations and right-click charts don't replace reporting. They eliminate the trip for the questions that never deserved a report in the first place — which is most of them.
The workarounds worth auditing this quarter
If I ran your platform team, this release would trigger a short, specific cleanup rather than a shrug:
- Inventory what you built to fake this. Custom UI, widgets, macros, client scripts — anything whose job description is "make the list do the thing the list wouldn't do." You probably know two off the top of your head. There are more.
- Test each one against the List Bundle behavior in a sub-prod instance before you assume parity. Some will map cleanly. Some won't. Find out which before you promise anyone a deletion.
- Delete with evidence, not vibes. A workaround that's been in prod for three years has users who depend on its exact quirks. Retire it deliberately, tell the people who use it, and keep the rollback.
- Go look at the AI flag if you're running Now Assist, and decide what your team does with the information now that it exists.
The uncomfortable part
The teams that will get the least out of this release are the ones that customized the hardest.
If you built a sophisticated workaround for pinned columns, you now have a maintained fork of a platform feature. It works, so nobody will want to touch it, so it will survive the upgrade, and in two years someone will be debugging why the out-of-box freeze behaves differently inside your custom list view. That is how customization debt actually accrues — not through one bad decision, but through a good decision that nobody revisited when the platform caught up.
"We already have that" is not the same sentence as "we don't need this." One of them costs you a maintenance burden forever.
Final word
Pick the ugliest list-related customization in your instance — the one whose original author has left the company. Open the Australia release notes next to it and ask one question: does the platform do this now?
If the answer is yes, you just found free capacity. Not a feature to evaluate. A thing to delete.
And if you'd rather not build the next one from scratch either, the community has already done some of the lifting: browse reusable flows and subflows on OnlyFlows — the boring, universal stuff shouldn't get built five hundred times.