Digitalization is influencing business models. At J.P. Morgan Asset Management, it is being used to advance the firm’s passive-investments business in Asia. That’s today. Tomorrow it is likely to be supporting a renewed push to engage directly with retail investors.
That’s according to Krishnan Sankaran, managing director and technologist. His background was in trading technology, and how he’s working on digital interactions in asset and wealth management.
Technology in banks has always been a moving target as large institutions maintain legacy systems and adopt new ones. Now that evolution is taking firms mobile, onto the cloud, and an I.T. architecture designed to be more oriented around services, not products or departments.
Then it’s down to artificial intelligence to transform these changes into a new way of doing business.
Developing A.I.
“The first thing you need in A.I. is data,” Sankaran said. How do firms model the data they already have, and how do they ensure that data is open to other parts of the firm, where appropriate?
“The second thing is, what questions do we need answered?” That requires technology teams not just working with data and code, but to be trained in asset management itself, so they understand how, say, natural language processing can read documents that plugs into the investing process. Increasingly firms are using A.I. to absorb and analyze structured (easy for a machine to read) and unstructured data sets.
The technology tools themselves are becoming commoditized, particularly as most are now open-sourced, meaning they are freely available to developers anywhere. “The libraries are now standard,” Sandkaran said. “Our ‘alpha’ is understanding finance and building a dictionary.” Then the firm can apply training data, and gradually its computers can process reports.
“There are plenty of applications,” he added.
Where to deploy
So when does J.P. Morgan let the machines start making investment calls? Is that an application ready to deploy?
“The investing process is the Holy Grail,” Sankaran said. “There are a lot of areas to go after for day-to-day use.”
Holy Grails tend never to be found. Is there more to A.I. than operational efficiency?
“Our industry is moving to passive investments, and that’s where automation helps. It involves a lot of data, in big volumes.” So the firm is looking to use A.I. to help portfolio managers involved in portfolio construction and smart-beta strategies, assembling and disassembling blocks of indices at various times – much more like an equities portfolio trading strategy, at least in its execution, than a long-term, buy-and-hold active fund.
What about for active portfolio managers? Sankaran mentions funneling alternative data sets, but it sounds like the emphasis is more on the passive side. Moreover, the passive business is an easier sell these days to retail investors – and digital technology, as it pushes further onto mobile and edge technologies, is more immediately relevant to consumer businesses than to institutional salespeople, who still rely on relationships.
“A lot of our product ends up in consumer apps,” Sankaran said. “There are a lot of mobile solutions out there, and our focus is on how consumers interact with us.”
So is J.P. Morgan A.M. doubling down on B2C business? Among global firms in Asia, J.P. and Fidelity are the only two shops that have made serious attempts at B2C business. Fidelity because of its brokerage licenses, and J.P. Morgan because of an acquisition in 2000 of Jardine Fleming, a major retail funds house.
But these efforts have never been successful – they probably would not stand on their own P&Ls – and other global fund managers regard B2C in Asia as the graveyard of ambitious CEOs. Banks control distribution around here.
“B2C is feasible from a technology point of view,” Sankaran said. “We have a retail platform in Hong Kong and in Taiwan, and we have new electronic trading tools. The question is, how do we make this bigger?”
