Dynamics 365 Plans Simpler Phase-In/Out Rules for Demand Planning in January 2027

The roadmap record lists the feature as in development, with estimated general availability in January 2027, for the Web platform in Worldwide Standard Multi-Tenant environments. That is a planning window, not a delivery guarantee: Microsoft explicitly describes its roadmap dates and descriptions as estimates subject to change.

What Microsoft plans to change​

The item outlines five areas of work:

• Shared selections: Preselect tables and fields at rule-group level and populate those selections across the group’s rules.
• Related-field autofill: When a key such as Product ID is included, fill associated values such as Product Name and Product Group.
• Standalone product introductions: Phase in a new product without identifying an old product.
• Time-series cleanup: Delete a dimension or dimension combination so it no longer appears in the time series.
• Potential AI assistance: Investigate rule creation through Copilot – Cowork.

These are roadmap-described capabilities, not instructions for functionality available today. In particular, the Copilot point is explicitly an investigation—not confirmation that AI-generated rules will ship with the redesign.

Why group-level defaults could matter​

Microsoft Learn documents phase-in/out as part of Demand planning. It lets planners simulate changes such as replacing products or opening a store or warehouse. Rules are organized into groups and applied through a Phase in/out step in a forecast or calculation model.

The documented authoring process involves configuring individual rules with dates, a rule type, a percentage factor and source-selection conditions. Phase-in rules also have destination-selection conditions.

That structure explains the potential benefit of group-level selections: when many rules use similar tables and fields, choosing that structure once could reduce repeated setup. This is a workflow inference, not a measured productivity claim. Microsoft has not supplied a time-saving benchmark or bulk-entry limit for this item.

The distinction matters. Easier high-volume entry does not, by itself, establish support for spreadsheet imports, a new bulk API or a particular number of rules per operation.

New products and deletion need more detail​

The proposed standalone product introduction deserves attention because today’s documented phase-in calculation draws a value from a selected source, multiplies it by a percentage and adds it to the selected destination. The existing workflow therefore makes the source-to-destination relationship explicit.

The roadmap does not explain which input would supply demand for a new product without an old-product reference. It would be premature to describe that as automatic forecasting for products with no history.

The deletion proposal also leaves important questions unanswered. Removing a dimension combination from a time series should not be interpreted as deleting product master records or historical transactions. The description does not establish the underlying deletion scope, recovery options or permissions.

The practical takeaway: both changes could expand useful planning scenarios, but implementation details will determine how administrators should govern them.

One existing behavior planners should check​

There is a useful consideration independent of the forthcoming authoring changes: where the Phase in/out step sits in a forecast model affects what it changes.

Microsoft Learn specifies that:

• Placing it before the Forecast step applies the adjustment to historical sales.
• Placing it after the Forecast step applies the adjustment to forecast results.

For teams assessing the redesign, reviewing that placement is a sensible preparation task. Faster rule entry is welcome; applying well-formed rules to the wrong stage of a model is still the wrong result, just reached more efficiently.

Microsoft’s documented model-design workflow also includes a Validate action that runs checks and reports issues before the model is saved. That provides an existing safeguard, although it should not replace reviewing whether the resulting forecast matches the intended business scenario.

The strongest supported promise here is less repetitive rule authoring. The broader product-introduction, time-series cleanup and possible Copilot capabilities remain worth watching—but should be evaluated against delivered behavior, not the ambition of a roadmap description.