Agentic AI in Aimplan: introducing the Aimplan AI Assistant and the Aimplan MCP server
Today we're releasing two new features that bring Agentic AI into Aimplan: the Aimplan AI Assistant, built into the administration portal, and the Aimplan MCP server. Both are available now in public preview for customers and partners.
Most planning and forecasting projects start the same way, with a clear idea of how things should work. There's an old spreadsheet process or a legacy system to replace, and a picture of what the new solution should look like. Getting from that picture to a working model is where the time goes. The new AI capabilities are built to shorten that path, and to suggest improvements to the process along the way.
The Aimplan AI Assistant: Agentic AI in Aimplan, built into the administration portal
You open the AI Assistant in the Aimplan administration portal and describe what you want to build in plain language. No other tools are needed.
The assistant reads the structure of your Power BI semantic model and checks what already exists in your Aimplan instance. Then it asks the questions that need answers before anything is proposed: which dimensions matter, at what granularity the plan should be set, and which measures and relationships are in play.
It draws on built-in skills based on best practice for configuring planning and reporting solutions, along with Aimplan's documentation and help center content. After working through your answers, it proposes a design covering both the Aimplan configuration and the changes needed in Power BI.
Nothing is built until you approve it
One principle shaped the whole feature: nothing is created or updated until you've reviewed and approved it. Every decision about what gets implemented, and how, stays yours. If you adjust the proposal, the assistant builds from your version, not its own.
Once approved, the assistant uses the tools in its Aimplan toolbox to create storage tables, measures, scenarios, and more. On the Power BI side, it can update the semantic model with the related tables, measures, and relationships. That means a new planning solution can go from idea to first working version considerably faster.
More than a designer
The AI Assistant also answers questions about Aimplan. When something doesn't behave as expected, you can use it to troubleshoot against what's actually in your Aimplan instance and Power BI model. And if you ask how an existing setup could be improved, it'll give you concrete suggestions to consider.
The Aimplan MCP server: the same capabilities in your own AI tools
A lot of Power BI development doesn't happen in a browser. It happens on your own machine, in Power BI Desktop or in PBIP projects stored in git. The Aimplan MCP server brings Aimplan into that workflow.
Connect Claude, ChatGPT, Copilot, or any other AI tool that supports the Model Context Protocol (MCP), and your tool gets the same capabilities as the Aimplan AI Assistant.
One AI tool, working across Aimplan and Power BI
Through the MCP server, your AI tool connects to Aimplan to configure storage tables, measures, scenarios, and more. At the same time, it can work directly with your local Power BI projects, updating semantic models and designing complete reports and planning solutions.
That means the whole solution, from the Aimplan configuration to the finished Power BI report, can be built with the AI tool you're already comfortable with.
Best practice included
The MCP server comes with best-practice skills for configuring Aimplan and the Aimplan custom visuals in Power BI. Your AI tool doesn't just get access; it gets guidance on how Aimplan solutions should be set up, which makes it possible to develop end-to-end planning and reporting solutions with Agentic AI.
Set up in about a minute
Connecting is simple. Copy the Aimplan MCP URL from the Aimplan portal and add it to your AI tool of choice.
Data, security, and availability
The first question most IT departments ask is what data the AI can access, and where it lives. Here's the answer up front.
AI Assistant: The assistant only has access to metadata, meaning the structure of your semantic model and your Aimplan instance. It never sees the numbers in them. Processing is handled by Anthropic's Claude. Every conversation is logged, and only the person who started a chat session can open it. The AI Assistant is switched off by default until an administrator turns it on.
MCP server: When you connect your own AI tool through the Aimplan MCP server, you control what that tool can access, the same way you do for any other MCP connection. The metadata-only design described above applies to the AI Assistant; with the MCP server, your AI tool's own settings and policies apply.
Both features are in public preview from the end of September 2026. Try them out and tell us what you think.
Power BI explains today. Aimplan shapes tomorrow.
Frequently Asked Questions
What is the Aimplan AI Assistant?
The Aimplan AI Assistant brings Agentic AI into the Aimplan administration portal. It helps you design, configure, and troubleshoot planning, forecasting, and reporting solutions in Aimplan and Power BI using plain language.
Does the AI Assistant make changes without my approval?
No. The assistant proposes a design and only creates or updates storage tables, measures, scenarios, or semantic model objects after you've reviewed and approved it.
Can the AI Assistant see my financial data?
No. The AI Assistant only accesses metadata about your semantic model and Aimplan instance, such as tables, measures, and relationships. It never accesses the rows of data in them.
Which AI model does the Aimplan AI Assistant use?
The AI Assistant uses Anthropic's Claude for processing. Conversations are logged, and only the user who started a session can open it.
What is the Aimplan MCP server?
The Aimplan MCP server lets AI tools that support the Model Context Protocol, such as Claude, ChatGPT, and Copilot, configure Aimplan and work with your local Power BI projects, using the same capabilities as the AI Assistant.
How do I get access?
Both features are in public preview for Aimplan customers and partners. An administrator enables the AI Assistant in the Aimplan portal, and the MCP server URL can be copied from the same portal.
