BNY, an investment and custodian bank in New York, is updating its AI tool, Eliza, named after Alexander Hamilton’s wife, to include multi-agent architecture to help its sales team make suitable recommendations for its corporate clients.
The bank created a lead recommendation agent and connected its various units so sales teams can access client and product information easily, thus providing valuable assistance to its sales teams and improving client engagement.
BNY has 13 agents that “negotiate with each other” to determine product recommendations based on different marketing segments, reducing the number of people salespeople need to consult with to provide customer recommendations.
The bank used Microsoft’s Autogen and LangChain to bring its AI agents to life, providing a solid framework giving the agents a set of rules to follow in their response protocol, making them more deterministic in their responses.
BNY plans to make further upgrades to Eliza, which serves as a blueprint for the bank’s further application of AI agents, with the next iteration aiming to be more intelligent.