Spendesk Brings Spend Data to Claude and Dust via New AI Protocol

Spendesk, the Paris-based spend and procurement platform, announced at Money 20/20 2024 a new Model Context Protocol (MCP) that lets finance teams retrieve spending data through AI assistants like Claude and Dust. The solution entered beta in June 2026 and marks a deliberate strategic choice: rather than competing with generative AI, the company is positioning itself as the infrastructure layer beneath AI workflows.

The MCP creates a secure bridge between AI assistants and Spendesk's API. Finance teams can ask natural language questions such as what their current cash position is or how marketing spending compares between quarters, and receive answers rooted in actual spend data within seconds. The protocol is deliberately limited in scope: it handles data retrieval and analysis only, keeping sensitive operations like approvals and payments under human control.

This architecture reflects a deeper shift in how finance teams work. For the last two years, generalist AI assistants have become embedded in day-to-day workflows. Rather than resist this, Spendesk recognised an opportunity. If finance teams were going to use Claude to draft emails and Dust to summarise documents, why not let those same tools answer questions about company spending?

Spendesk's own data shows the scale of this transition. AI tools now represent 38 per cent of all technology spending recorded on its platform, up from less than 1 per cent four years ago. That is growth of 50-fold in four years. Year-on-year, AI tool spending through Spendesk's platform accelerated 281 per cent in early 2026 alone. The company analysed 24.3 million transactions worth €4.17 billion across its user base from January 2022 to May 2026 to quantify the trend.

Quentin Vigneau, Spendesk's Chief Product Officer, said the spending pattern demanded a rethink of how finance software works. "This spending trend redefines what finance software needs to be," he said. "In this context, success is no longer determined by feature breadth, but being the infrastructure layer that lets finance teams work with the tools they choose."

The move reflects a broader pattern in fintech. Most spend management software was built around the assumption that finance teams would use that vendor's interface. Spendesk is accepting that assumption no longer holds. Teams will use multiple tools. The vendor's job is to be the reliable data source underneath those tools, not the application layer on top.

Spendesk was founded in 2016 by Rodolphe Ardant, Guilhem Bellion, Jordane Giuly and Quentin Nickmans. The platform now serves 5,000 companies across more than 35 countries, handling corporate cards, invoice payments, subscription management and broader company spending. It has raised €312 million across multiple funding rounds, with the largest being a €100 million Series C round in 2022 led by General Atlantic and Tiger Global. The company reached profitability in the first quarter of 2025.

Spendesk appointed Axel Demazy as CEO in October 2024, bringing new leadership as the company matures. The timing of the MCP launch suggests the company is accelerating product innovation beyond its core spend management features. Customers can request early access to the protocol directly through the platform, positioning it as a controlled rollout rather than a general release.

For European finance teams already purchasing AI tools at accelerating rates, Spendesk's move hints at a future where every financial dataset becomes queryable through conversational interfaces. The question is no longer whether finance software will integrate with AI, but how deeply and on what terms.