Asset & Wealth Management
Private clients: big assets. Big data?
Private banks are just beginning to play with robo advisors and artificial-intelligence tools.
As this week’s survey has shown, private banks are using digital technology, to improve efficiencies and to generate better client experiences. Tech has become a competitive differentiator.
But what about the most innovative applications of data? Consumer banks with mass-affluent investors are now mining user data to construct portfolios, to seek higher investment returns, to enhance product development, and to personalize offerings.
Private banks catering to ultra high-net-worth individuals and families are dipping a toe in these waters, but in different ways, and with caveats.
“A.I. and big data could be the differentiator in product innovation,” said Kam Shing Kwang, head of private banking in Asia at J.P. Morgan. Like many private banks, her firm has other divisions: investment banking, capital markets, asset management, prime brokerage. “We can leverage the data we already have,” she said.
Differences over data
This makes sense for banks serving tycoons at the helm of extensive businesses.
To a degree, such data aggregation can also play a role at monoline private banks such as Pictet Wealth Management, which is a pure-play asset manager for wealthy clients. “Of course we analyze how clients use our tools, so we can refine the information we send to them,” said Claude Haberer, CEO for Asia. “It’s important to fine-tune our offering.”
However, the diversity of data and what it can do means that banks are taking different approaches.
“We’re not likely to use A.I. for a robo advisor,” said the Asia CEO for ultra-HNW at a U.S. bank. “But using machine learning to understand trends in terms of what clients want, and to data mine for prospecting new clients? That’s at an early phase.”
Serving the client
To derive insights from client data first means a bank needs a killer app that clients like to use. This has to be more than just digitizing paper statements; it has to help clients make decisions. And then the bank has to find a way to plug this into its legacy I.T., so it can pull client data from various mainframes traditionally dedicated to product lines, not customer profiles.
One bank that has taken steps down this path is Credit Suisse, which has rolled out its “Digital Private Bank”. The firm’s head of private banking for Asia Pacific, Francois Monnet, says putting big-data analytics to work will require bigger changes in the bank’s operating model, however.
“We need to invest in context management,” he said, “and mining the user activity online.” For example, can banks provide their relationship managers with video and other storytelling tools, to make sense of the data? That’s where A.I. and client data comes in.
“The last mile of optimization is not a new tool, but a new way to service the client,” Monnet said.
One Hong Kong company providing A.I. tools to private banks is MioTech. Jason Tu, its co-founder and CEO, says banks are using data to provide RMs with better ways to work with clients – who themselves are better informed thanks to the internet.
“The clients know more than the R.M.s now,” Tu said. “It’s harder for RMs to make the case for a particular investment idea. So they need a data story and they need visualization tools.”
The clients know more than the R.M.s nowJason Tu, MioTech
Companies like MioTech crawl the web for data, including traditional financial feeds, plus trends in news and social media, and then other online sources and apps, which can get very specific. Then they map that to their proprietary “knowledge graphs” that connect events or topics with specific companies or people.
It’s the same work a human analyst would do, but instead of taking weeks to collate information, the A.I. can do it in less than an hour.
Enter the robo-advisor
Another area that is creeping into private banks is robo advice. Few banks are actually looking to apply this to their ultra HNW clients. But Goldman Sachs and J.P. Morgan are using robo in the U.S. for their lower-tier HNW clients, in the $2 million to $10 million range of assets.
In Asia, these firms are more focused on just the ultra client, usually $25 million up to billions of assets. Their trading desks are already using A.I., however, both for quant strategies and for other alpha-generating investment strategies. And private bankers in Asia are looking at robo, if not yet deploying it.
Robo’s role is less obvious, but it will play a partKam Shing Kwang, J.P. Morgan
Ricardo Wenzel, director at KPMG, a consultancy, says private banks are exploring robo-advisory portfolio management tools and digital channels.
“An important consideration here is client-management efficiencies, without compromising client experience,” he said. Lower-value clients require lower-touch servicing as investments are mostly around vanilla, listed products. This is easier to automate.
The value of robo for ultra-wealthy clients is more about allowing them to self-direct trades if they prefer to do so.
The future value of the R.M.
But robo will not replace R.M.s (although they are likely to resist it). There is no way a computer can structure a complex asset allocation, advise on restructuring a company, or set up a philanthropic foundation. The same probably goes for structured products and other complex products. And in times of financial crisis, rich people will probably demand human assurances.
And then there’s the biggest obstacle of all: how to integrate a robo-advisor into a bank’s legacy systems.
But for daily transactions and investment ideas? That’s already easy to automate.
“Robo’s role is less obvious for ultra high-net-worth clients,” said Kwang. “But it will play a part.”
Private banks are not yet jumping into big-data analytics or robo. But the fact that they’re studying these things, and starting to talk about them, is already a change.
There will be limits to what a private bank can or should do with data. But how exceptional is the private-banking model? And does this mean they will remain aloof from the big-data and A.I. trends now engulfing other parts of financial services?
First, bankers say, they serve relatively few clients, and therefore have few data points they could use.
DigFin does not believe this will remain the case. The customers may be few, but they own or influence big companies, which have extensive financial activities. Big data will play a huge role across these touchpoints. And those private banks that are able to aggregate this will have an even bigger edge.
The second argument against big data is that the most important client information is private. It must be given in person, often for compliance reasons, and is therefore off the table. That may be true, but DigFin’s view is there are now so many new sources of data out there – on people, on companies, on things that impact their portfolios – that it is naïve to think that this won’t impact the ultra rich.
In fact, the opposite is more likely: the ultra rich have lots of assets in lots of places. The balance of data that can be found online will become far greater than information gaps due to family secrets or complexities of estates.
The third argument against rushing into A.I. is that private banks offer bespoke, sophisticated service, so there’s no need (or desire) to use tech to personalize offerings, as happens in the mass consumer space.
DigFin believes this is true, but also notes that the financial products and the investment ideas on offer are pretty commoditized, even for ultras.
What remains bespoke are things like estate and tax planning, generation management, legal advice, and softer things like building art collections or endowing a philanthropy. These are true differentiators. But investments? The only sustainable edge there is the delivery of content – the U.X., the data, the algos – not the content itself. (Sorry, active fund managers.)
The final hedge against private banks’ using A.I. and big data is trust. Private banking is above all a trust business. Technology is a tool that can deliver efficiencies, convenience and better service. How it gets deployed, especially with regard to treating customer data, will matter most of all.