Billtrust MCP Server links AR data to Claude and Microsoft Copilot

Billtrust said it has launched the Billtrust MCP Server, a connector that lets finance teams ask questions about live invoice-to-cash data from inside Microsoft Copilot and Claude. The company said the product gives users structured answers from Billtrust data without requiring them to sign into a separate system or assemble reports from multiple applications.

Billtrust described the launch as the first accounts receivable platform connection of its kind using the Model Context Protocol, or MCP. Accounts receivable, often shortened to AR, covers the invoices a company has issued but not yet collected; invoice-to-cash refers to the process from billing a customer through payment and cash application.

What is the Billtrust MCP Server?

The Billtrust MCP Server is a read-only connection between Billtrust’s AR platform and AI assistants including Claude and Microsoft Copilot, according to the company. MCP is a protocol that allows AI systems to connect to external tools and data sources in a structured way, so an assistant can retrieve relevant business information for a user’s request.

Billtrust said the server gives access across four areas: invoicing, payments, cash application and AR analytics. In practice, a finance chief could ask for a summary of AR risk before a quarter closes, while an AR manager could ask which accounts are becoming late and receive a ranked view by outstanding balance and days past due, according to examples given by the company.

The product is designed to put AR information into the AI workspace where finance staff are already working. Lee An Schommer, Billtrust’s chief product officer, said the company’s payments intelligence network has been developed over 25 years and that the launch changes where customers can use that insight.

“Finance teams shouldn’t have to leave the AI workspace where they’re already making decisions,” Schommer said. “The Billtrust MCP Server puts AR data where it’s more accessible, so the answer is always one question away.”

How Billtrust says the connector uses payments data

Billtrust said answers can draw on connected systems such as ERP, CRM or financial planning and analysis data, alongside the company’s AR intelligence. That intelligence is trained on anonymized payment data from 13 million buyers and $150 billion in annual payments processed, according to Billtrust.

The company said the first release is limited to structured, read-only access. A second phase planned for later in 2026 is expected to add a fully embedded Billtrust user interface within the AI tools and a pre-built prompt library. Billtrust also said it plans to extend the server beyond read-only access so finance teams can take actions inside their AI tool and turn repeated tasks into automated workflows.

Kevin Permenter, research director at IDC, said the lack of integration has held back AI adoption in finance operations. He said AR data is a predictive financial signal because it can show movement, risk and potential operational stress, and described Billtrust’s MCP Server as a step toward making that data visible inside AI tools used for finance decisions.

Billtrust also outlined its security model for the connector. The company said each customer’s data remains in that customer’s own environment and is not exposed to other users. AI tools receive structured answers rather than raw data access, each query is logged with an audit trail, and setup is configured once through the identity provider already used by the customer, according to Billtrust.

This story draws on original reporting from Finextra Research.