The Model Context Protocol (MCP) has gained significant traction since its launch in November, with many users adopting it as a standard in the industry. However, regulated industries, particularly financial institutions, are hesitant to join the MCP and Agent2Agent (A2A) bandwagon due to compliance and safety concerns.
While regulated companies have started experimenting with AI agents, such as banks and hospitals, the integration process has been slow and meticulous to ensure compliance. Sean Neville, co-founder of Catena Labs, highlights the importance of establishing fundamental building blocks for interoperability and communication in the AI ecosystem.
Financial institutions face challenges in integrating AI models and agents into their existing risk frameworks. Greg Jacobi from Salesforce notes that regulated entities are wary of the non-deterministic outcomes produced by AI agents, causing concerns about quality control and risk management. Despite these challenges, there is potential for financial institutions to embrace MCP and A2A in the future as the technology evolves.
