MAIN CONTENT
From digital assistants to agents that take action
At the Global Financial Leaders’ Investment Summit hosted by the Hong Kong Monetary Authority, bankers and investment executives examined how AI is changing finance. An agent differs from a conventional chatbot by taking a stated objective, selecting tools and performing a sequence of tasks. That difference matters: an inaccurate summary on a screen is not equivalent to an unauthorized transaction that has already been executed.
According to remarks reported by the South China Morning Post, State Street’s Manoj Bohra said the group had identified more than 350 AI use cases, up from about 150 last year. Thousands of colleagues already exchange ideas with AI during their work. The next challenge, he suggested, is deciding whether parts of those workflows can be delegated to agents operating within explicit boundaries and with identifiable human supervision.
The increased tally of applications does not mean all have been made fully autonomous. Document analysis, internal assistance and systems authorized to initiate financial transfers carry markedly different risk profiles. Counting deployments is useful, but assessing their significance requires information on data access, approvals, testing and the consequences of an error.
















