harris richard 200

Artificial intelligence (AI) is beginning to transform the way credit and fraud decisions are managed by the global finance industry.

The transformation is driven by an ‘eruption’ in the level of data being generated, coupled with computer and software advances that can filter and analyze this data to provide insights and trigger decisions.

AI is already influencing all areas of business – but should prove especially useful for the finance industry. This will, however, require a significant change in the way lenders approach their customer base, according to Richard Harris, Head of International Operations at Feedzai, which specializes in machine learning services for businesses, speaking at White Clarke Group’s Auto Captives Summit in November.

Harris argues that - with computing power now a fraction of what it was a decade ago - every lender can integrate artificial intelligence and machine learning into their business to provide highly personalized services that focus on individuals, not customer groups.

He told the Summit: “In the past, what we did was build scorecards or we built rules. We made decisions about people by putting them in a bucket. Such as: where are you from? Where did you go to school? How long have you been in employment? How much is your house worth? All these criteria of stability and probability were used to make decisions about individuals.

The time has come to stop putting people in boxes. You’ve got all this processing power; it’s cheap. You’ve got machine learning algorithms which can look down to a very fine-grained level of detail at hundreds of different factors. You don’t need to group people anymore.

“Let’s treat people as individuals! Let’s profile every single person as an individual rather than a classification!”

Richard Harris goes on to argue that technology is now so advanced, and affordable, that customers may only need to provide a couple of key pieces of data for the computer to complete entire forms and guide them through a complex decision-making process.

Systems can automatically link to a wide range of sources, including social media, to provide a much better understanding of each customer.

They can also use machine learning to understand subtle indicators of fraud.

Harris added: “All the analytics data, the clickstream, how you interact with that webpage, is a huge predictor of whether you’re a good guy or a bad guy.”

Artificial intelligence specifically, machine learning can also be used as an accurate predictor of future behaviour, for areas such as fraud but also for identifying trigger points for purchasing and the likely requirements of an individual.

This isn’t the sole domain of large tech companies like Amazon, Facebook and Netflix anymore and as it becomes more extensively used throughout the industry, all suppliers and service providers will have to adapt to a more personalizsed offering.

“The future is not building a model,” Harris stressed, “The future is having the power to run the platform yourself and have control of the model within your own world and say, what do I want to happen? That’s where the change is going to be. Own the artificial intelligence and have the control of it yourself.”