Finom Integrates with Claude and ChatGPT via Model Context Protocol for SME Banking

European SME financial platform Finom has launched Finom MCP, integrating business accounts directly with AI assistants like Claude and ChatGPT. By adopting the open Model Context Protocol, the firm allows founders to query live transaction data and invoices using natural language, removing the friction of switching between AI workspaces and banking applications.

What was announced

Finom MCP is a connection built on the open Model Context Protocol (MCP) designed for SMEs and entrepreneurs. It allows users to interact with their financial data from within third-party AI assistants. At launch, the tool provides read-only access, enabling customers to ask questions about transactions, wallets, invoices, cards, and general accounting. For instance, a user could request the assistant to "analyse all my spendings for 2026" or "list transactions over €500 last month" directly inside the interface they already use for drafting and planning.

The integration is included in all Finom plans across Europe and is rolling out pan-EU starting today. Security is a central pillar of the rollout; the connection is initiated by the customer through their AI assistant's settings and authenticated via a Finom-hosted screen. This ensures no login credentials are shared with the AI provider. Access is strictly scoped to the user’s existing role within the Finom platform, and the connection can be revoked at any time. While the current phase is focused on data retrieval and a morning briefing of recent activity, the company has outlined a roadmap that includes preparing everyday actions for user confirmation and expanding the range of product areas that plug into the connection over time.

"Every serious conversation about AI and business finance is now happening inside an assistant, where founders already draft, plan and analyse. Their account should be reachable from there. We are not the first to public MCP and we will not pretend to be. We’re starting with a first set of read-only access, with write capabilities to follow, all built on an open standard rather than a walled garden."

Kristjan Kaar, CPO of Finom

The companies involved

Finom is a European financial platform serving more than 300,000 businesses and entrepreneurs. The company has positioned itself as a comprehensive hub for SME financial management, combining banking, invoicing, and accounting services into a single interface. Operating out of the Netherlands, Finom focuses on the digital-first needs of modern founders, frequently integrating advanced technology to streamline back-office operations.

The firm operates in a highly competitive European neobanking landscape, vying for market share against other business-focused fintechs. Its strategy has increasingly leaned into artificial intelligence, led by Chief AI Officer Ivo Dimitrov and Chief Product Officer Kristjan Kaar. By adopting the Model Context Protocol rather than developing a proprietary, closed integration, Finom is aligning itself with an emerging open standard for AI-to-tool communication. This approach reflects a broader market trend where fintechs are seeking to become the "financial operating system" for businesses by embedding their services where users already spend their time, such as in Large Language Model (LLM) interfaces.

What FF News has reported before

FF News has tracked Finom’s rapid expansion and product diversification through 26 previous reports. In June 2026, the platform launched a native AI assistant to automate SME financial management, a precursor to today's external integration. This followed a partnership with Porters aimed at automating back-office operations. Beyond software, the company has expanded its physical product line, including the Metal Premium Business Card. Finom has also sought to increase its brand visibility through high-profile sports marketing, recently signing a multi-year deal to become the Official B2B Financial Platform of Juventus FC.

What this means

This move signals a shift in how business banking is consumed, moving away from standalone apps toward "headless" banking where the primary interface is an LLM. By adopting MCP, Finom is betting on an open ecosystem rather than trying to lock users into a proprietary AI environment. This puts pressure on traditional business banks that still rely on siloed, manual reporting tools. The challenge for the sector now lies in moving from "read" to "write" access; while querying data is low-risk, allowing AI to prepare transactions raises significant questions about liability and the robustness of "human-in-the-loop" verification in automated workflows.