Within conventional wisdom, Artificial Intelligence (A/I) is still associated with science fiction – namely robots and cyborgs, eventually turning against humans.

Aside from one or two limited experiments – you may remember IBM’s Deep Blue or more recently Google’s AlphaGo, its use or perceived use had been quite low. But now think of the latest developments of personal assistants and driverless cars as Artificial Intelligence applications, and you realize things are rapidly changing across different industries. From process automatization tools, natural language understanding to cognitive automation, Artificial Intelligence initiatives are gaining more traction.

Data Acquisition, Processing and Modelling

To understand why, we need to go back to its definitions. In computer science, the field of A/I research defines itself as the study of “intelligent agents”: any device that perceives its environment and takes actions that maximize its chance of success at some goal(2). Artificial Intelligence truly materializes, when the solution has “the ability to learn, understand and think in a logical way(3) “when the solution can adapt to its environment beyond its original design”. As such A/I solutions must embody two key capabilities, each with varying degrees of complexity: data acquisition & processing capabilities, and modelling capabilities.

New Game Changers?

While a huge amount of research has been conducted since the 1940s on machine learning/neural network techniques, implementations were originally secluded to research papers and prototyping at universities. All that changes thanks to the latest technological shifts: decreasing hardware costs, cloud technologies, big data, and natural language processing to name a few. These are the game changers, effectively opening, as an IBM CEO stated four years ago, “the third era of computing technology […], where machine learning meets big data […] and decisions will be made using predictive analytics and data”(4).

Taking advantage of these advancements would have been beyond the reach of most Sell-Side and Buy-Side players just 10 years ago. However technology behemoths and Fintech are now making it available ondemand. To name a few: Microsoft Cognitive Toolkit and Azure; Google TensorFlow; Amazon AI services and AWS Deep Learning AMI, Salesforce Einstein and IBM Watson.

A/I And Asset & Wealth Management

Possible applications of A/I in asset and wealth management are endless and impact the entire value chain. Imagine sentient, real time optimization of sales and marketing interactions and client services , predictive market modeling based on instantaneous processing of teraflops of data, self-creating exchanges with complete price transparency. It is not unrealistic to think that research analysts will have their own bots using real time big data to do a majority of their market and financial analysis, and even assembly of reports and recommendations to the CIO.

Laborious trade reconciliation could be a thing of the past, with AI bots on both ends of the transaction across back, middle and front office doing it all and alerting humans only in rare cases of disputes. If (or when) quantum computing becomes pervasive all trading, investment and distribution decisions could be done through bots leaving asset and wealth managers free to focus more on client relationships and business strategy. Of course, A/I initiatives are still a big bet to make. Business cases are hard to sell due to still relatively high entry costs, steep learning curves, uncertain synergies/gains, and questions about data and privacy regulations. And will all this lead to higher market volatility or less? What new profit mechanisms might arise?

Share this post