Bots for banks, and smarter finance

While futurists have dreamt of intelligent machines for decades, if not centuries, the mainstream has been resistant to change. Discarding these dreams as lofty ideas, entire industries have remained stuck in bygone eras with legacy systems that no longer add value to customers. No doubt, we’ve come a long way in a single century — from the Industrial Revolution to the Information Age — but there’s no denying we’ve fallen well short of some targets. Cities in space. Jetpacks for all. Good customer service. Unfortunately, it’s the economic engines of the world, our oldest industries, that have struggled the most with embracing change. But their customers are calling for change, and the calls are getting louder and louder.

Introducing, artificial intelligence (AI) — we all should be pleased to meet its acquaintance. Even the banks are trying to get around it.


Just the beginning


Bank chatbots and robo-advice are just the beginning. They pale in comparison to other AI-powered technology that will fuel financial services in the future. Australia’s largest bank, Commonwealth Bank, only rolled out its chatbot at the start of 2018. Affectionately named Ceba, the bank expects it will ultimately understand 500,000 ways that customers might ask for 500 different banking activities. Then there’s RoboChat, developed between IBM Watson and UBank, which specifically responds to questions on Australian home loans. There’s a theme here. In most cases, legacy financial systems have adopted this technology from other places — Credit Union Australia (CUA) has also partnered with Flamingo, ANZ has teamed up with Nuance, Westpac is working with local startup Hyper Anna.

Whether this will be enough to retain customers, or they will lose customers to the innovators, is a question that will be answered in the future.


More than bots for banks


AI is clearly making its way into all areas of traditional financial services, from transfers to trading and anywhere the cash flows in-between. On the contrary, from the get-go, cryptocurrency products and platforms were built with AI front of mind. Are banks taking a page out of the crypto-playbook? It certainly seems so. In the UK, incumbent investment bank Barclays has recently teamed up with AI startup Simudyne to simulate possible futures. This will provide another safeguard in risk modelling, potentially better preparing Barclays analysts for all sorts of financial events, from minor market corrections to black swan situations. Risk modelling is nothing new in banking, but this is a new spin on an old thing — and a problem that countless startups have been working on solving for quite some time. Where traditional financial analysis relies on regression-based models, through artificial intelligence, Simudyne delivers an ‘agent-based approach’. Basically, the old way means scanning piles of historical data for common themes, even though past performance isn’t an indicator of future performance. Alternatively, the new ‘agent-based approach’ can take limited, formerly uncorrelated data sets and still come up with possible futures. Puzzling questions like “how should interest rates be set for price stability” and “what happens to unemployment in a recession” may be rendered trivial. If this succeeds, it will be able to run simulations of entire economies millions of times under all sorts of conditions — and land on accurate solutions.

This is simulation as a service, powered by AI. And, as always, this is technology created on the fringes and only after adopted by legacy institutions. That’s the law of the land.

Leave a Reply