A few weeks ago, our CEO and co-founder Matthieu answered questions from the newspaper Le Soir on the influence of artificial intelligence on wealth management. Here are the best bits for those who missed the print edition.
Yes, we use AI on many levels at Eaysyvest. Already since our launch in 2016, we have been using a “machine learning” algorithm to predict the future performance of our portfolios. It is for this reason, among others, that we are called "robo-advisor", even if we devote the majority of our resources to direct interactions with our customers. More recently, we have completely adopted the new tools that are chatGPT, Midjourney or Github Copilot, in our IT developments and our marketing and administrative activities.
Currently, we see chatGPT primarily as an individual coach for each company employee. Every employee gains in productivity. For example, we estimate that every IT developer has been developing for six months 30% faster. So, with a team of three developers, it is as if we had hired an additional developer practically for free!
Financial regulation has two objectives: to protect the consumer investor and to ensure confidence in the stability of the financial system. For example, if a wealth manager does not provide adequate financial advice to a client, its managers may be subject to sanctions, including criminal ones and imprisonment. But today, chatGPT often fantasizes when asked financial questions: when he doesn't know, he invents, or he never gives the same answer twice. The regulations are right to protect the investor and to sanction institutions that do not go in this direction. As such, we are not currently ready to allow our clients to communicate directly with chatGPT as part of their portfolio management at Easyvest: the risk of errors and subsequent penalties is still too great.
Given the exponential nature of the development of AI, apprehending its evolution at 10 years is a perilous exercise for the human mind, which thinks linearly. The evolution is likely to be much faster than we imagine: remember that chatGPT is only six months old and yet has already radically changed the way we carry out our IT developments or build our marketing campaigns. At the same time, history shows that we tend to overestimate how quickly new technology is adopted and underestimate the long-term impact it can have on society as a whole.
I see two possible scenarios. First, it is not difficult to imagine that in 10 years, a financial institution will be able to manage the same number of customers as today with two or three times less human resources, which obviously raises fundamental questions of social protection and employment. If a financial institution arrives at this state before all the others, the competitive advantage that it will derive from it in terms of the cost of capital will be such that it will have the financial means necessary to become a monopoly. On the other hand, it is not impossible that this leaner cost structure will allow managers to reduce fees for their clients.
When talking about large sums of money or financial advice on a heritage acquired by the sweat of his brow over a long career, we prefer to speak and deal with a human. So far, no bot has demonstrated that it can convey the trust needed to completely dispense with a human advisor. If I think that AI makes it possible to envisage a back office almost without human beings, I do not believe in it at the front office level. Money is about trust, emotions and empathy: all elements best shared from human to human.