Sophia Leung, chief information officer for Asia Pacific, says the bank engaged with a local A.I. provider this year to develop Asian-language chatbots, seeing the need for cognitive automation for salespeople and analysts.
“These are information-intensive jobs and we leverage on some of the existing technologies to enhance our ability in advising clients,” Leung said. In Hong Kong in particular the bank’s asset-management arm works with clients in multiple languages. Although the bank already uses chatbots in the U.S., these solutions were not suited to Asian languages.
“Globally, we are a wholesale bank,” Leung said. “But a lot of emerging technology is now digital, coming from customer demands, and people’s expectations are rising.”
The bank is also leveraging machine learning to make client experience unique and intuitive with personalization. J.P. Morgan recently partnered with Amazon in the U.S. to provide institutional clients an option to access research via Alexa, its digital personal assistant. And the bank is testing other features, like providing prices on bonds or swaps.
The bank is working with Professor Pascale Fung and the Center for Artificial Intelligence Research at Hong Kong University of Science and Technology on a research project. Fung is known for her work in spoken language systems, natural language processing and empathetic human-robot interaction.
It is also working with external partners that augment the bank’s internal innovations, Leung said. “A lot of our strategies are hybrid now. We don’t want to get locked into a single vendor.”
DigFin understands the Asian-language proof of concept is with Clare.ai, a Hong Kong-based fintech launched in 2016 by Bianca Ho and Ken Yeung. It now covers ten Asian languages plus English. The company works with multinational companies, including asset managers, telcos and insurers. Enterprises are adopting local-language chatbots to cut costs in client servicing as well as support relationship-managers, traders and others to communicate on messaging apps such as WeChat; the company’s website says it is now testing its chatbots on WhatsApp.
A.I. tools create new challenges. Robot process automation is straightforward, but artificial intelligence needs to be explainable, especially given the heavy regulation around banking. So Leung’s team is proceeding on a trial basis with various applications. Some applications are simple, like using chatbots to give customers instant access to account information, but even in these cases, fintech partners need to be able to work with the bank at an enterprise level.