Using the Dynamics 365 ERP Analytics MCP for Client Finance Questions

The Dynamics 365 ERP Analytics Model Context Protocol (MCP) server allows conversational AI tools to ask questions directly of a company's financial and operational analytics data. For fractional CFOs and accounting advisory firms, this provides a secure, auditable way to get fast, evidence-based answers from a client's live ERP system using natural language. It is important to note that, according to Microsoft documentation, the Dynamics 365 ERP Analytics MCP is a preview feature and its capabilities may change.

The business problem: Manual client reporting from Dynamics 365

Fractional CFOs and their teams often manage a portfolio of clients, many of whom use different instances or configurations of Dynamics 365 Finance and Operations. Answering a client's ad-hoc question—about expense variances, cash flow drivers, or accounts receivable aging—typically involves a time-consuming and manual process:

1. Receive the question via email or during a meeting.
2. Log in to the client's specific Dynamics 365 environment.
3. Navigate to the correct module and run one or more standard reports.
4. Export the data to Microsoft Excel.
5. Manipulate the spreadsheet—pivoting, filtering, and summarizing—to find the specific answer.
6. Draft a response, often copying and pasting the findings.

This workflow is inefficient, doesn't scale across a growing client base, and introduces the risk of manual errors. It also creates a delay between a client asking a question and the advisor providing a data-backed answer, reducing the impact of the advice.

Illustrative example: Reviewing a client's cash flow drivers

A fractional CFO is in a virtual board meeting with a client's management team. The CEO asks, "We felt a cash crunch last quarter. Can you quickly tell me what the top three drivers of our cash usage were?"

Instead of deferring the question, the advisor opens their firm's approved AI assistant, which is connected to the client's Dynamics 365 ERP Analytics MCP. They type a prompt:

Because the MCP provides the AI with tools to query the client's Business Performance Analytics module, it can access the necessary data while respecting the advisor's user permissions. The AI assistant returns a concise answer:

The advisor can now lead a strategic discussion based on live data, all without leaving the meeting or manually running a single report. This workflow elevates the advisor's role from a report-puller to a real-time strategic partner. For more complex forecasting, this can be a starting point for a deeper analysis using tools for MCP for cash-flow forecasting.

Use case readiness checklist

Before implementing an MCP-based workflow for a client on Dynamics 365, use this checklist to assess readiness:

• The client uses Microsoft Dynamics 365 Finance, Supply Chain Management, or another compatible Finance and Operations application.
• The "Business Performance Analytics" module is configured in the client's D365 instance, as this is a prerequisite for the ERP Analytics MCP.
• The advisory firm has an enterprise AI assistant (e.g., Claude, ChatGPT Enterprise) with appropriate security and data privacy controls.
• User roles and data access permissions are correctly defined within D365 for the advisory team members who will be using the MCP. The MCP will enforce these same permissions.
• The client has provided explicit, written authorization for the advisory firm to connect an AI assistant to their live ERP data for specified purposes.
• The firm's client engagement letter has been updated to include clauses on AI tool usage, data handling, and access protocols.
• The finance team has identified a pilot set of 3-5 key financial questions or reports to test the MCP connection (e.g., P&L variance, AR aging, top vendor spend).

Prerequisites and limitations

Connecting an AI assistant to a client's ERP is a powerful capability that comes with important considerations.

Prerequisites:

• Software Version: The client must be on a version of Dynamics 365 Finance and Operations that supports Business Performance Analytics and the ERP Analytics MCP.
• Feature Enablement: The Business Performance Analytics feature must be enabled and configured within the D365 environment.
• Authentication: A secure connection requires setting up OAuth 2.0, allowing the AI assistant's MCP server to authenticate as a specific, permissioned user.

Limitations:

• Preview Status: The Dynamics 365 ERP Analytics MCP is a preview feature. Firms should carefully evaluate its stability and functionality before relying on it for critical client work. Features and availability are subject to change.
• Read-Only Analytics: This MCP is designed for querying and analyzing data. It provides read-only access and does not support write-back capabilities like creating purchase orders or posting journal entries.
• Data Scope: Access is limited to the data available within the Business Performance Analytics layer. It is not a direct connection to all raw transactional tables in the database.
• No Data Rights: Using an MCP to access client data for advisory work does not grant the advisory firm any ownership, permission to sell, or rights to use that data for AI training. Data licensing is a completely separate process that requires explicit authorization from the client company's decision-makers.

Questions to ask your software provider or implementation team

1. What specific version of Dynamics 365 is required to use the ERP Analytics MCP, and is it considered generally available or still in preview?
2. How does the MCP enforce the data access policies and user permissions already configured in our client's Dynamics 365 instance?
3. What are the technical steps and required permissions to configure Business Performance Analytics as a prerequisite for using the analytics MCP?
4. What kind of audit logs are generated when an AI assistant queries data through the MCP? Can we review who asked what, from which system, and when?
5. Does the MCP provide access to raw transactional tables, or does it work exclusively through the pre-built analytics and reporting layers?
6. What are the potential costs associated with this workflow, including any API call charges from Dynamics 365 or usage fees from the AI model? See our guide to ERP MCP costs for more detail.

Next step with SourceX

As you use tools like the Model Context Protocol to streamline your advisory workflows, you are working directly with the structured, high-quality operational data inside your clients' systems. This same data—representing real-world business transactions and outcomes—can be extremely valuable to AI labs and data buyers for training next-generation models.

SourceX helps companies license their anonymized operational data. As a fractional CFO, you are in a unique position to identify clients who may be a good fit. The process is simple: you make a permissioned introduction to an authorized decision-maker at a qualifying company. SourceX handles the technical evaluation, data fit assessment, and commercial contracting.

You can use our free, no-login Company Fit Checker to quickly screen clients in your portfolio. If an introduction you make leads to a signed deal, your firm receives 25% of the platform fees SourceX collects, up to $100,000 per referred company. This is a share of SourceX's fee, separate from the licensing revenue paid to your client.

Sources

• Anthropic finance agents (May 5 2026)
• Intuit QuickBooks Online MCP (Current official repository)
• NetSuite AI Connector Service FAQ (Current Oracle docs)
• Dynamics 365 ERP MCP (Current Microsoft docs)
• Dynamics ERP Analytics MCP (Preview)
• Power BI MCP overview (Current Microsoft docs)

Vendor capabilities change. Check current official documentation before relying on any product detail.