For financial decision-makers, optimizing operational efficiency and accuracy directly impacts the bottom line. Microsoft Dynamics 365 Business Central has moved beyond simple automation into the era of AI Agents. Unlike traditional tools, these agents operate autonomously in the background to manage end-to-end workflows. This executive article provides an updated overview of these transformative tools, highlighting their current release status, capabilities, and the underlying technology driving their integration.
A foundational principle across all these AI features is human oversight. While these agents perform the bulk of the work, users (referred to as “Agent Supervisors”) are kept in the loop for review and approval, ensuring compliance with established business rules.
Key Business Central AI Agents and Copilot Features
1. Sales Order Agent
Status: General Availability (GA) The Business Central Sales Order Agent fully automates the sales order-taking process—from initial inquiry to final confirmation. As of late 2025, this feature is officially out of preview and available for production environments in supported regions (including the US, UK, Canada, and Australia).
- Autonomous Monitoring: The agent monitors a shared mailbox and runs a background task every 20 seconds to identify new requests.
- Multi-Turn Clarification: If an email is vague, the agent engages in professional correspondence to clarify items or quantities, considering inventory and customer preferences.
- Availability & Promising: Beyond simple searches, the agent can use “Capable-to-Promise” (CTP) logic to calculate realistic shipment dates for out-of-stock items.
- Human-in-the-Loop: The agent drafts the quote and the reply email, but a human supervisor must click “Confirm” before any communication is sent to the customer.
2. Payables Agent
Status: General Availability (GA) / Advanced Features in Preview The Business Central Payables Agent automates vendor invoice processing. While core invoice registration is now Generally Available, advanced features like Purchase Order (PO) Matching and Confidence-Level Prioritization are entering public preview in early 2026.
- Data Extraction & E-Documents: Incoming PDF invoices are processed via Azure Document Intelligence. Each document creates an entry in the Inbound E-Documents table, providing a full audit trail.
- Strict Processing Guardrails: To ensure performance, the agent is currently limited to 100 emails per day. PDF attachments must be under 5MB and a maximum of 10 pages.
- Intelligent Creation: If a vendor is not found, the agent can draft a new vendor card using OCR data. For security, these new vendors are created in a “Blocked” status until a human verifies the details.
- Reasoning Transparency: The agent provides “hints” on the purchase document draft, explaining why it chose specific G/L accounts or matched a specific vendor.
3. Bank Account Reconciliation Assist with Copilot
Status: General Availability (GA) This feature supplements the traditional “automatch” rules in Business Central to handle complex financial discrepancies.
- Secondary Pass Matching: After standard rules run, Copilot performs a second pass using AI to match transactions based on fuzzy logic (e.g., matching a lump sum payment to multiple smaller invoices).
- Post Difference to G/L: Copilot can now suggest entries for bank-side transactions that don’t yet exist in the ledger (like service fees or interest). It suggests the most likely G/L account (e.g., “Bank Charges”) based on the transaction narrative.
- Language Consistency: For optimal results, Microsoft recommends that G/L account names and bank statement descriptions remain in the same language.
The New Frontier: Model Context Protocol (MCP)
A major technical milestone in the 2025 Release Wave 2 is the introduction of the Model Context Protocol (MCP) Server for Business Central.
What is MCP? MCP is an open standard that enables a seamless, “pluggable” connection between AI models (like those in Copilot Studio or Azure OpenAI) and your Business Central data.
Why it matters for Executives:
- Extensibility: Your IT team or partner can now build custom agents that “talk” to Business Central using a standardized protocol. This means you can create highly specialized agents (e.g., a “Collections Agent” or a “Warranty Specialist”) that understand your specific business logic without complex custom coding.
- Contextual Intelligence: MCP allows the AI to understand the context of your ERP data—knowing the difference between a “Customer” and a “Vendor” automatically—leading to more accurate responses and fewer “hallucinations.”
- Future-Proofing: By adopting the MCP standard, Microsoft ensures that Business Central can integrate with future generations of AI models as they are released.
Master the Shift to AI Staffing: > Managing AI effectively requires more than just a license; it requires a leadership change. Read our guide on why managing digital coworkers like humans will change everything with Microsoft Agent 365 to learn how to transition your team to this new collaborative model.
General Considerations for Adoption
- Consumption-Based Billing: These agents utilize Copilot Credits. Administrators can monitor usage and set limits via the “Copilot & AI Capabilities” page in the Business Central Admin Center.
- Permissions and Security: Agents do not have “god-mode” access. They operate within specific Agent Permission Sets and Profiles assigned by the administrator.
- Infrastructure: Both the Sales Order and Payables agents require a dedicated Microsoft 365 Shared Mailbox. To prevent conflicts, users should not access these mailboxes directly via Outlook.
Conclusion
Microsoft Dynamics 365 Business Central has transitioned from “AI-assisted” to “Agent-led” operations. With the General Availability of the Sales Order and Payables Agents, and the architectural leap provided by MCP, organizations can now automate high-volume, repetitive tasks with unprecedented reliability. Financial leaders should prioritize establishing a “billing and oversight” framework to begin scaling these agents across their finance and sales departments.
Don’t just automate—innovate with custom AI. While Business Central’s native agents handle the basics, the real competitive advantage lies in custom-built agents that know your industry. Take our AI Agent Business Impact Assessment to bridge the gap between “standard features” and a fully integrated AI workforce designed specifically for your organization.
