It’s important to consider what functions AI and other technologies are most effective at managing. This includes processing huge volumes of data and generating intelligent output — whether it be portfolio recommendations, investment commentaries or banking transactions. Human intelligence is limited in terms of the volume of data our brains are capable of processing. But how intelligent is machine-generated output? Let’s look at the use of AI in the financial services sector in the three key areas mentioned above:
1. Portfolio Recommendations. Robo-advisors leverage AI technology to optimize their portfolio investments. They also have used technology to make the client experience much better by providing digital reporting that meets compliance requirements, but also offers added value, such as investment planning and “what if” scenarios that help their clients plan their financial future.
Although this package of services will be sufficient for simple, and typically smaller, portfolios, robo-advice is not sufficient for large, complex financial relationships in which highly skilled professionals assist with financial planning, wills and estates planning and tax-optimization strategies.
The successful wealth manager of the future will leverage AI and machine learning to support investment recommendations, but will need to apply human intelligence qualities such as creativity, emotional sensitivity and ethical judgement to maintain and build their client relationships. Even top AI experts agree that it’s unlikely machines will ever replace these human qualities. In fact, according to a recent report from PricewaterhouseCoopers LLP, 26% of wealth- and asset-management firms globally use AI to inform, but not make, their next “big decisions.”
2. Generating Intelligent Output. The emergence of natural language generation (NLG), a form of AI, is taking place in investment-management firms’ back offices. NLG technologies are effective at ingesting large amounts of investment data and generating what are referred to as “base” investment commentaries. Although the technology is ground-breaking and delivers incredible value, it requires humans to complete the “last mile” by using judgement, intuition and strategic thinking to complete the commentary.
A great example of NLG’s success is a case study involving the Associated Press (AP). Before leveraging NLG technology, AP’s writers were able to produce 300 corporate earning stories each quarter. However, by leveraging NLG, AP now produces 4,400 quarterly earnings stories — a 12-fold increase over their manual efforts.
Still, NLG has its limitations. Machines will never be able to add strategic content, such as the portfolio manager’s recommendations on market positioning and why her portfolio is positioned for success. Experts anticipate that NLG will be able to deliver 75%-80% of the investment commentary and human expertise will be required to complete the final and perhaps most important component: strategic recommendations.
3. Banking Transactions. The digitization of retail banking is taking place as many banking functions that humans previously managed have been replaced by technologies that include AI. Managing account balances, applying for and approving loans, and transferring funds are obvious automation opportunities. The productivity and client convenience benefits of automating these functions are very compelling.
However, there are limitations. As the next wave of technology and AI emerges, a recent report from global consulting firm McKinsey & Co. estimates that “machines will do up to 10% to 25% of work across bank functions.” The limitations of AI and other forms of digital automation are the same in retail banking as they are in other financial functions mentioned earlier — the human touch.
There’s a fear that despite the massive convenience of digital banking, eliminating this human touch will risk brand loyalty. Specifically, loyalty toward the bank that manages all aspects of clients’ finances stems from the advice and service from individuals with names — not the technology used to process basic transactions.
Overall, the evidence suggests that AI will continue to evolve as an aid to human interactions rather than a replacement of these relationships — especially in financial services. In fact, AI will just make these interactions that much better — both for the financial services sector and clients alike.