Sygnum Bank tests AI agent for live transactions
Zurich/Singapore - The digital asset banking group Sygnum has become the first Swiss bank to test Artificial Intelligence (AI) agents for live-on-chain transactions. The pilot project is based on a principle in which AI complements, but does not replace, human decision-making. As such, customers retain control at all times.
(CONNECT) Sygnum, a global digital asset banking group headquartered in Zurich and Singapore, has become the first regulated bank in Switzerland to test AI agents for live on-chain transactions, as detailed in a statement. The agent uses client-issued plain text instructions to execute multi-step live transactions on a blockchain Mainnet. The AI agent independently plans the steps, checks relevant smart contracts and flags potential transaction risks before it presents each transaction to the customer for approval.
The system is based around a “human-in-the-loop” design, which constitutes a core principle of the AI@Sygnum development team. AI-supported transactions are therefore only signed via the client’s own digital wallet installed on their own device. In this way, private keys remain on the customer’s own device and under their control at all times.
“The key challenge is doing this in a way that preserves – and even enhances – bank-grade consent, custody and trust”, comments Thomas Frei, Head of AI and Data Analytics and AI@Sygnum lead at Sygnum Bank, in the statement. “That is what we set out to solve, and what this pilot demonstrates: That a regulated Swiss bank can provide clients with the speed, convenience and accessibility of agent-driven execution, without ever giving up control of their assets”, he concludes.
The pilot was developed using the open standard Model Context Protocol (MCP) server and utilizes Claude from Anthropic as its underlying AI model. The AI agent is not yet available to customers, as its deployment remains subject to full regulatory, compliance and security reviews and approvals. ce/ja
