When your bank knows your future

Computers are getting very good at predicting life events such as retirement, says IBM's Boxley Llewellyn. Picture / supplied

Major financial institutions are using artificial intelligence to predict big events in their customers' lives.

Boxley Llewellyn, a US executive at IBM specialising in predictive analysis for banking, wealth management and insurance firms, said artificial intelligence was getting better at crystal ball gazing.

"Artificial intelligence is starting now to become a good predictor of the future," Llewellyn said at the INFINZ - Institute of Finance Professionals New Zealand - conference this month.

"For instance, if you took all of the banking data for a small business and looked at all behaviours of the individuals involved, you could start to predict their good moments and their bad; we can predict life events very accurately through behaviours and data in your transactions systems."

Given that machines were learning and getting smarter, the financial sector needed to think about where that capability can best be applied.

"We're starting to talk about systems that every day reach out, and in the case of banking, looking for customers who are unhappy and looking for investment ideas or risk portfolios that are interesting and those systems have to learn," he said.

Llewellyn said that Texas-based financial giant USAA found that by looking at their data, they could pick up on big events in a business' or person's life before they happened.

"Somewhere behind that, the data says: something's happening. The data is looking like a rumbling before the big earthquake, it is the disturbance in the force. We helped them create this work and took nine different life events and said can we predict those based on the rumbling? It may be a positive event like a retirement where we can get more of their investable income ... one hundred per cent of the time we could predict them very logically and very correctly," he said.

While there was trepidation about robots replacing bankers, Llewellyn said a better path was to use AI for simple customer tasks - like changing passwords - and let the experts deal with more complex parts of the job.

AI could also augment a worker's skills rather than replace staff altogether, he said.

Brazilian banking giant Bradesco, for example, was using AI to feed information to customer service staff to help them deal with clients.

Andy Lark, chief growth officer at Ansarada, told the conference the "robotic revolution powered by AI is well under way and happening".

Lark talked about how his firm was using AI to help clients who are raising capital on equity markets or looking at mergers and acquisitions.

"We can now present in a tender or a acquisition or a divestiture or some kind of M&A activity, we can now predict with 94 per cent probability after six to seven days who the probable winner of the bid will be and the approximate value they will likely win at based on their behaviour," he said.

"That's what AI does; it's the end of not knowing. There's no guessing around 'do you think someone is going to buy us?' I can tell you very precisely whether or not you're going to be bought, sold, or raise money or complete a tender."