Wealth managers slowly switching on to rewards of AI and machine learning

While machine learning and artificial intelligence (AI) based systems are transforming many industries, there is often a feeling that asset and wealth management have been left behind. But this is about to change, believe technology providers.

Many ‘in-memory’ computing technologies, which can enable real-time machine learning and AI-based systems are being quietly developed by finance industry players, says Terry Erisman, VP of marketing at US tech provider GridGain Systems. So far the moves have been low-profile, and restricted to improving business efficiency, rather than consumer-facing applications. High-speed automated asset trading and real-time regulatory compliance are examples of the innovation, typically along the lines of “turning Big Data into real-time insights or Fast Data,” says Mr Erisman. The beauty of this technology, he says, is that new solutions can simply be plugged into the legacy plumbing rather than building a system from scratch.

“High speed trading is clearly growing in popularity and usage,” he says. “They come to us because our solution can slip in between their existing database and application layers, so it is easy to install and realise the speed and scale benefits of in-memory computing, without replacing their existing data layer architecture.”

New players

A greater number of non-traditional players will enter the industry – currently in a state of flux – during the next five years, to service needs of new-generation investors, believes Mr Erisman. “This will cause increased competition and investors will begin moving assets between firms.”

This means both traditional banks and newer, tech-led boutique firms will need to find ways to distinguish themselves, providing better service and lower costs than rivals, with wealth managers also subject to an increasing degree of more intrusive regulatory scrutiny, he says. “They will need to operate in a 24/7 world and provide access for investors across many platforms and technologies such as mobile and even wearables and IoT (internet of things) devices.”

One of the problems with machine learning is that the investment management community as a whole sees it as a tool for back-testing methodologies, but not for day-to-day incorporation in investment techniques, where they prefer a rule-based approach, says Gregory Michaelson, director at DataRobot Labs. A more clandestine cohort of enthusiasts is striding ahead with the development of innovative solutions, while much of the industry is being left behind.

“There is a small population of computer scientists and quants that are deeply engaged in building highly complex, highly customised solutions to specific problems,” says Mr Michaelson, while the greater part of the asset management community has deep knowledge of trading, but little background in modern machine learning methods.

“The main barrier for data scientists is the very steep learning curve when it comes to markets, while the main problem for traders is lack of experience with machine learning,” he suggests. While the possible applications of machine learning are there, they are not so obvious to the financial community. “To me, that means there is huge potential to generate alpha using these approaches,” says Mr Michaelson. “But it’s going to take the creative thinking of asset managers, combined with advanced tools that enable them to do more technical work.”

Culture shift

The problem, he suggests, is a cultural one of data and techno-phobia. “Even the most sophisticated mainstream institutions that are trying to leverage AI are not meeting their goals. Most firms still view machine learning and artificial intelligence as a technology that requires a PhD-level expertise and months or years of work to implement.”

A cultural change is necessary in financial institutions, he suggests, with “everyone from the most junior analyst to the most senior fund manager” needing to understand the opportunities in AI, fuelling a culture in which the company tries to execute as on as many opportunities as possible, rather than building a handful of solutions. “It’s about solving as many technical problems as possible using advanced tools to automate most of the technical work,” says Mr Michaelson.

The main lesson financial firms can learn from the innovators in Silicon Valley and elsewhere on the US west coast is speed to market, he believes. “It’s certainly the case that large financial institutions cannot seem to build as fast or effectively as fintech companies. Solving this problem can certainly translate to real value.”

Mr Michaelson also believes robotics, combined with developing sophisticated trading algorithms, can help portfolio managers to beat the market. “The beauty of automation is that it’s making it possible for the best traders to use the best algorithms to test and evaluate their ideas in a robust way,” he says.

But other technology providers believe the focus should be more on the client servicing side. “Our experience is that wealth firms are now beginning to look seriously at how machine learning and AI could be utilised, particularly from a client engagement perspective,” says Robert Roome, global head of technology at Wealth Dynamix, who believes a historical nervousness of trusting cloud-based systems has limited wealth firms’ ability to develop innovative platforms common to other industries.

Storage of client data has also been a problem, with poor quality client information often held across a number of siloes, negatively impacting the effectiveness of technology tools. The impetus for change is now being provided by a new breed of tech-savvy start-ups. “Traditional wealth firms that fail to adopt this new thinking run the risk of being subject to unflattering comparisons, not just against the other services which clients use, but also relative to robo-advisers operating in their own space,” says Mr Roome.

In orders to capture more of this start-up mindset, involving more rapid development and delivery of products and constant communication with clients, larger firms need to either look to partner with fintech and technology firms or recruit their own teams of specialists to work in incubators and innovation labs, he says.

The fast pace of regulatory encroachment will provide another stimulus to innovation. “With an increase of regulation, including the EU’s MiFID II and GDPR (General Data Protection Regulation) coming next year, we see the need for firms to rapidly adapt as these rules evolve and grow,” says Mr Roome, who has seen an increasing focus on ‘regtech’ providers in the market to help gain scale in meeting these challenges. This label covers a huge variety of solutions, from use of machine learning for identification of high-risk client conversations, to building of digital platforms that help clients directly enter and refresh their KYC (Know Your Customer) information. “While regulation is still a key driver for IT spend, firms that find ways to deliver both a better customer experience and an enhanced compliance framework at the same time will stand to gain most,” believes Mr Roome.

While he agrees that many traditional firms have been caught on the back foot during the last five years, as they have failed to adequately respond to digitisation of financial services and the emergence of robo-advice in particular, he expects to see them fighting back for the remainder of the decade, either through partnerships, or building their own tech centres of excellence.

“This will inevitably lead to more sophisticated automated propositions for the lower end of the market, but also to the adoption of sophisticated tools for traditional investment managers to service those clients who still want and need a more personalised approach,” says Mr Roome.

Examples of these innovations could include digital personal assistants to help clients plan their time and complete administrative tasks, and machine learning used to pro-actively predict client needs. The future of artificial intelligence in wealth management, predicts Mr Roome, is certainly a bright one, laying to rest the ghosts of the analogue past.