Edition: May/July 2018
Editorials

ARTIFICIAL INTELLIGENCE

A help to do better

Continuing our series on the impact of fintech on asset managers, Fatima Vawda of 27four looks forward to a ‘force multiplier’.

Knowledge is a coin with two sides. The factual side consists of numbers and data that are easy to find, copy and reproduce. The other side is less tangible. It exists in the skill of the craftsman with years of experience or the doctor who has seen thousands of patients. It cannot be converted to facts and figures, yet it performs a crucial role.

On this latter side, artificial intelligence (AI) steps in. Intelligence can be viewed as the speed and efficiency of an agent’s adaptability; how quickly it can solve a problem at the lowest cost, avoiding mistakes, wastes and repeats.

AI as a concept is not complex. In its basic form it uses computer power to look for patterns in old data and apply it to new data, giving the illusion of intelligence. It is the sheer volume of processed factual data that reveals patterns and allows an algorithm to mimic know-how, experience and judgment.

When you take a picture of your family with a cell phone, the circle around the faces is enabled by AI. An algorithm identifies features that reveal faces. The more features it can identify, the better it is at “recognizing”.

Despite its apparent simplicity, AI can be a powerful tool. It scans the world of data, looking for patterns and using brute force to make discoveries. Where in the past we had to rely on little bits of knowledge spread across hundreds of people for an invention to emerge, a single computer program sitting in the cloud can source all the information on a topic and then try millions of combinations of this information until it finds a solution that makes statistical sense.

Vawda . . . small things add up
Vawda . . . small things
add up

More importantly, AI can look at data that humans cannot readily process. It can see connections. Using satellite imagery to count cars outside shopping malls, for example, it can provide a financial analysis of real-estate companies. Or by measuring the shadow length of the roofs of each oil storage tank in the globe to estimate current oil reserves, it can also look back in time to estimate usage and thereby forecast oil prices.

As fund managers, we are not using AI technology to generate elaborate new ideas but rather as a force multiplier. We are looking for an information edge, by converting data into information for better investment decision making.

We can present a program with all of the financial statements of companies that went bankrupt, for example, to see whether we can develop better indicators of a company’s financial health. We can use this information to avoid troubled companies, or provide feedback to managers who will hopefully take corrective action.

Like many industries, fund management is being changed by AI. There will be a few big developments that make headlines. But as in many other industries, the real impact will be in the countless small things AI will help us do better, faster and cheaper. They’ll add up to produce significant benefits.