'AI in robo-advice is coming, but just not yet', Alex Ypsilanti, Chief Executive Officer, Quantifeed
"While AI promises to change the future of investing, the robo-advisor industry isn’t quite there yet in terms of integrating this new technology into its investment models. In fact, it may not do so for a while."
Could artificial intelligence, AI, outperform humans when it comes to investing? This is a question on the minds of many in the finance sector right now. But while AI promises to change the future of investing, the robo-advisor industry isn’t quite there yet in terms of integrating this new technology into its investment models. In fact, it may not do so for a while.
Some providers in Asia Pacific use AI in their services, but at this stage it all seems experimental and nobody is touching the investment models. On the contrary, the majority of the industry at the moment is working on expanding the current service models which are based on the core purpose of robo-advice: the scaling and the automation of wealth management.
There are good reasons too for the industry to continue to focus on the existing models before trying to integrate AI into robo-advice services.
Firstly, the investment models and principles used for the past three decades still work and are likely to continue to work in the future. Technological advances can refine them and bring more computer power to the table, but the core concepts remain the same. In other words, AI may not be used in any creative or intelligent form to try to outperform humans in investment strategy any time soon.
Another reason is that there’s a whole set of innovations just starting to be digested by the broader investment community. Even though risk questionnaires, low cost investments like ETFs, and immediate market access through online brokers have existed for a few years, it’s only with today’s digital wealth management services that all these elements have been integrated onto a single platform. So for the most part, people are only benefiting from these innovations now.
This doesn’t mean that the current robo-advice services don’t have anything innovative to offer; quite the contrary. Robo-advice solutions bring a new element to the table in wealth management services, namely the removal of the human bias. This is a very significant improvement.
Robo-advice technologies -as we know them today- can make decisions that are based on pre-set rules and never be tempted to opt for products with better commissions for themselves. Robo doesn’t get greedy or scared. So, for financial institutions looking to develop a digital wealth management offering, the current models of robo-advice are not only about scaling a service, but also about improving the quality of a service and building the trust of clients.
There are also unexploited advantages in the current models from a compliance point of view. With robo and digital advice, companies can establish parameters in their services, making it clear and objective in regards to where the boundaries are for their clients’ portfolios. In contrast, with a human advisor, there are internally approved product lists and models, but unfortunately there are always ways to reach outside these parameters. With robo, the implementation of those parameters is auditable and transparent.
Room for AI development in specific areas
AI is certainly coming, but we’re still years away from a broad use of it. As an industry, we are only now crossing the bridge towards a more personalised generation of robo-advisors.
The first robos had rules for standard risk-profiles and investment goals. The next generation can personalise the service by drilling down to some of the requirements of a person that are outside the scope of a risk-profile. Personalised models take historical data from a client’s portfolio and can learn about the behaviour of an investor. These new models are able to tell with what type of allocations a client may be more comfortable; and they can put to the test what’s been said in a risk-profiling form. With these robos, two people with the same goals and risk profile can each have specific portfolios crafted for them.
And although AI may be integrated with robo-advice, companies are more likely to use this technology to improve user interaction and user experience rather than creating new investment models. Chatbots, computers powered by AI, can help to improve the interaction with customers in a friendly and engaging way and even provide intelligent guessing based on the clients’ parameters.
Finally, there’s an important question that financial institutions engaging with AI should ask themselves: “How do I explain AI to the regulator?” Real AI may be unpredictable and opaque. On the contrary, robo-advice is simpler from a regulatory perspective, given that is essentially an automated and scalable service based on pre-set rules. Having a type of intelligence that creates a portfolio based on unwritten rules is a risky business.
For now, robo-advice services powered by existing investment models will continue to serve customers well. The machine outperforms humans in terms of discipline, transparency and scale. But the investment models and the rules are still very human.