With Fabric IQ Plan, Microsoft is introducing a new feature within Microsoft Fabric. Fabric Plan focuses on planning, write-back, and business workflows directly on data. Specifically, this means:
- A single environment for planning, analysis, and reporting
- Works directly on the same data in Fabric (OneLake / SQL / semantic models)
- No need for separate tools such as Excel-based planning or external CPM systems
Fabric Plan offers the ability to replace Excel-based budgeting and forecasting. In addition, data entry and data corrections can also be integrated as part of broader use cases.
Key Components of Fabric Plan
Planning Sheets
Planning Sheets can be used for budgeting, forecasting, and scenario planning. It works with dimensions such as time, department, and product, allowing planning data to be recorded in a structured and flexible manner. In addition, it supports rules and validations to ensure data quality. The database can be populated from a Semantic Model or (once) from existing Excel files. This can replace typical Excel budget models.
How do the Planning Sheets work?
You can choose to use the actual figures from the Semantic Model as the basis for the planning data, but in this example, the budget data is already available in Excel and is imported into Fabric Plan for further processing. In this case, “Date” and “LabelCode” are purely technical fields, intended to be linked later to the Date dimension and Label dimension in my Semantic Model.
In a Planning Sheet, it looks like this, and I've added a Forecast column as an extra—it's a copy of the Budget column, which I've increased by 2%.
What’s interesting is that various scenarios can be created based on the initial plan. These can be positive or negative scenarios, or, for example, a new version that is updated quarterly based on actual results. For a new scenario, you can specify—at both the aggregate and cell levels—the extent to which the forecast should increase or decrease, after which the changes are automatically rolled up to a higher level. In the example below, the projections for Label Services were adjusted downward during the summer break, while those for the other divisions were increased.
Ultimately, these various scenarios can also be used in the report and compared with the actual results.
PowerTable Sheets
Another interesting feature in Fabric Plan is PowerTable Sheets. This is essentially a no-code mini-app for:
- data entry
- Master Data Management
- workflows
A good alternative is to replace Excel spreadsheets and set up “business-driven dimensions” for Semantic Models. As an example, a PowerTable has been created in which a custom sales region can be defined for each country.
This is done through a user interface, where various validation rules can be associated with a column (database field). In the example below, the Country field is marked as unique and required, and the input form option has been applied.
Authorization and Audit
For each Planning Sheet (and even for each scenario), read or write permissions can be granted to users so they can view or edit the Planning Sheet. If desired, a different user can be designated to perform the actual “writeback” to the database when additional oversight is required.
Authorization can be set individually for each PowerTable Sheet. In addition, the system records who made which changes, and comprehensive track changes can be configured to retrieve a record's previous values.
Fabric SQL Database
Before you can start using Fabric Plan, you must first create a Fabric SQL database. This is where the data for Fabric Plan is stored.
For each Planning Sheet and PowerTable Sheet, you can specify which Fabric SQL database, schema, and table it should be linked to. Depending on user permissions, you can run the desired queries on these tables and combine them with other data sources.
In the case of a Planning Sheet, in addition to the dimensions and measured values, the scenario is also displayed, along with information about who made changes at the record level and when.
Relationship with BI and Power BI Semantic Models
Traditionally, BI has been primarily read-only. With Fabric Plan, it becomes translytical (combining analysis and transactions). This creates a closed loop for users:
Analyze → adjust → save → analyze again
The question then is: Are changes immediately visible in Power BI reports?
This depends on the chosen architecture.
Fabric Plan stores write-back data in a Fabric SQL database. When the semantic model is (partially) linked to this data via DirectQuery (for example, only the planning data), changes are visible in real time. However, this can impact performance and limit certain functionalities.
An alternative is near real-time. In this scenario, for example, a user initiates an action—such as clicking a button—after which a pipeline runs in the background that reprocesses only the scheduling data.
There are two ways to set this up. One is to reload (part of) the Semantic Model in Import mode. The other, when using Direct Lake, is to synchronize the data to the Lakehouse so that it is immediately available for reporting.
AI and the Bigger Picture
Fabric Plan is part of the Fabric IQ stack, which includes Ontology, Graph, and AI Agents, among other components.
When an AI agent is linked to a semantic model, it is not enough to simply know the data—the agent must also understand the business context. That is why the AI agent is linked to the Ontology, which defines business terms (such as Customer, Revenue, Order) and rules. In addition, an AI agent must also be able to reason about data. This is done via the Graph, which captures relationships between entities (for example, Order → Shipment → Sensor → Incident). This enables the agent to establish connections across multiple data sources and explain insights. Fabric Plan rounds out this system by adding an operational layer. This provides AI with more context, enabling it to better support decision-making and propose targeted actions.
Fabric Plan vs. Other Planning Tools
Microsoft also offers planning functionality within its own Dynamics suite. In addition, there are specialized planning solutions on the market, such as Jedox and Anaplan. These tools feature a comprehensive calculation engine and support complex processes such as consolidation and top-down and bottom-up planning.
Fabric Plan does not (yet) offer this advanced functionality, but it does bring planning to the data platform. This allows planning data to be enriched directly and combined with BI and AI applications.
For organizations that still do their planning in Excel, this offers a logical first step toward centralizing planning and better integrating it with their data and analytics environment.
Conclusion
Although Fabric Plan is still in preview, its Planning Sheets and PowerTable features offer an alternative to traditional, Excel-based solutions. It enables organizations to centralize planning processes while improving governance, control, and data quality within workflows.
For organizations that already use specialized planning solutions, planning data will often be transferred to Fabric via ETL processes. In practice, however, there are often still separate Excel spreadsheets and manual updates. This actually presents opportunities to structure and standardize these fragmented processes within Fabric.
AI is playing an increasingly important role within Fabric, and by incorporating planning data, AI can not only perform analyses on historical data but also provide context by combining historical data with planning information, thereby helping organizations make better decisions and propose targeted actions.
