BI Book AI - FAQ

If a user has access to a report, and BI Book AI is enabled in the BI Book environment, they can use BI Book AI with that report. At the moment, you cannot selectively allow or deny BI Book AI access for specific users only. However, you can limit the amount of tokens per user role to control and manage AI usage.

Yes. As an Admin, you can enable or disable individual AI models to match your company’s preferences, compliance requirements, or cost policies.

It can be provided that you disable Claude (Anthropic) and Gemini as model providers in BI Book AI - Settings page. This is because even though we have Enterprise agreements with all of these, Azure OpenAI is currently the only one that can ensure that all prompts are processed and stored within EU. At the time of writing this, 8/24/2026, Microsoft’s terms say Prompts are stored for 30 days for Abuse monitoring purposes. See the terms of service for the most recent information.

As for Anthropic and Google Gemini, they process data in GDPR-compliant regions.

Then when it comes to BI Books solution itself, all runs and data is stored and processed in EU.

Please read more at BI Book AI - BI Book.

You can access the core security, privacy, and compliance documentation for BI Book AI from the following documents:

• BI Book | Terms of Service
  Describes the general terms governing the use of BI Book and BI Book AI, including responsibilities, rights, and usage conditions.

• BI Book AI Addendum
  Specifies AI-related terms, including how BI Book AI handles data, model usage, and AI-specific responsibilities.

• Data Processing Agreement (DPA)
  Covers data processing roles, subprocessors, security measures, and GDPR-related commitments for BI Book and BI Book AI.

No, your prompts, AI outputs, and business data are not used to train generative AI models. They are also not shared with other customers and not used by the AI provider to improve their models. Moreover, BI Book does not use customers’ or authorized users’ personal data to train AI models. However, BI Book may use anonymised and aggregated platform usage data (which does not directly or indirectly identify you or your users) to improve the BI Book AI services.

Yes, you can. We have released BI Book AI MCP server, that allows you to connect your own AI tools to the BI Book AI. Read more at MCPs - BI Book

BI Book AI supports models from three leading AI providers: Anthropic (Claude), Google (Gemini), and OpenAI (GPT). This gives you flexibility to work with the latest generation of large language models, with GPT enabled by default and Claude and Gemini available as optional alternatives.

Yes. You can define and control monthly credit/token limits for users in your company.

When a user reaches their AI credit/token limit, the chat application gets locked. They cannot continue using AI features until the credits are reset or increased by an Admin.

Yes, you can. However, BI Book AI requires that your BI Book report's semantic model be published to a Power BI workspace provided by BI Book.

No, at the moment, BI Book AI features are limited to copying and exporting data from the chat window.

Yes. BI Book AI supports alerts and AI-driven recommendations, but this feature is currently in a limited pilot phase and not yet available for general use. General availability is planned towards the end of 2026. If you’re interested in early access and helping shape how this feature develops, please contact your Sales Responsible Person to join our pilot user program.

You can use BI Book AI to read any data inside your reports, even your planning data, and you it to create forecasts for you. BI Book AI not yet fully integrated with planning, so you need to copy paste or otherwise input the estimates yourself into the planning tool.

Yes. You can upload or reference your process handbook in the AI instructions/context so BI Book AI learns your internal vocabulary, follows your established processes, and tailors its answers to your organization’s practices.

You can copy any data output directly to your clipboard using the copy button and paste it into Excel or any other tool of your choice.

Only an admin can do this. Have your admin head to the BI Book AI section in BI Book and open Settings. There, they'll find the MCP URLs ready to drop into Claude, VS Code, or whatever environment you're connecting from.

✅ AI models
Microsoft Fabric Copilot uses Azure OpenAI and doesn’t let you switch to other model families (like Gemini or Anthropic) within the Copilot experience.
BI Book AI lets you choose and switch between multiple model providers.

✅ Reliance on the semantic model
Fabric Copilot quality depends heavily on how well your semantic model is built and annotated in Power BI.
With BI Book AI, our experts handle the modeling and optimization, so you can get high‑quality AI assistance without deep semantic model work.

✅ Licensing & capacity
Fabric Copilot typically requires user-specific licenses and sufficient Fabric capacity + AI compute for meaningful use.
With BI Book AI, running on top of BI Book, you don’t need separate Copilot licenses or to manage Fabric capacities for AI features.

BI Book AI shows the underlying DAX queries behind every answer, so you can see exactly how a result was calculated and verify it against your data. That said, like any AI system, it can make mistakes — we recommend reviewing the generated queries before acting on critical business decisions.