For as long as most Dynamics 365 teams can remember, the year had two fixed points. Twice a year a release plan arrived: a long document listing, app by app, what Microsoft intended to ship over the next six months. You read it, mapped the features to your processes, decided what to test and what to switch on, and moved on.
That document is not coming back.
What changed
On 25 August Microsoft announced that it is, in its own words, "retiring the twice-yearly release wave 1 and release wave 2 model in favor of continuous publishing." From September, new Dynamics 365, Power Platform and Dataverse capabilities are disclosed on the AI at Work roadmap as they become committed, rather than gathered into a plan every six months. There is no 2026 release wave 2 plan.
Existing items with a public preview or general availability date from 1 June 2026 onward are moving across between September and November. The Release Planner retires by 15 November. The 2026 wave 1 plan is still on Microsoft Learn, now marked as archived and no longer maintained.
The new roadmap is one place for Microsoft's business applications, Microsoft 365 and Copilot together. You can filter it by product, export it, subscribe to it by RSS and share a saved view. Items move through three states: in development, rolling out, launched.
What it does not change
Most of the coverage has run this as "Microsoft ends release waves", which is half right. Microsoft is explicit that the change "doesn't change how our products are built, released, or deployed", and that products with established release schedules will keep following them.
So your Finance and Supply Chain service updates still arrive on their calendar. Message Center is still where Microsoft tells you what is about to land in your tenant. Feature management in finance and operations still decides what is switched on in your environment. None of the machinery moved.
What it does change
What goes is the ritual. The release plan was never just information; it was a trigger. Twice a year it turned up, somebody was asked to read it, and a planning conversation followed whether anyone had diarised one or not.
Continuous publishing removes the trigger. Items now appear when they are committed, in any week, one at a time. That is better information and a worse habit. Without a deliberate review, the first anyone hears of a change is a user asking why a screen looks different.
The second change is quieter. One roadmap now spans Dynamics 365, Power Platform, Dataverse, Microsoft 365 and Copilot. For a finance and operations team, a good share of what will actually affect you is filed under Copilot or Power Platform, not under Finance. A filter on your app name alone will miss it.
One caution
"In development" is not a date. The release plans carried the same warning — features could slip or not ship at all — but a single roadmap with tidy statuses makes it easier to treat a line item as a commitment. Do not put an in-development item into a project plan, and do not promise it to the business.
What we would do first
- Build the filter this week, before the Planner goes. Your Dynamics 365 apps, plus the Power Platform, Dataverse and Copilot items that touch them. Save it as a shareable view so the whole team is looking at the same list.
- Send it somewhere people already look. Subscribe to the RSS feed from the channel your Dynamics 365 team works in. A feed nobody opens is the same as no feed.
- Give it an owner and half an hour a month. Microsoft itself suggests a monthly or quarterly review. Triage each new item into one of three piles: ignore, test, or plan.
- Keep testing tied to the updates, not the roadmap. Your regression pack should run on the service update calendar and on what Message Center says is coming, both of which are exactly where they were.
None of this is difficult. It is the kind of small, unglamorous routine that falls into the gap between Microsoft and the project team, which is why it so often does not happen. It is part of how we run Dynamics 365 estates under application management. If you would rather not own the filter and the monthly triage yourselves, that is the conversation to have.