Artificial intelligence promises much, but we’ve only seen the tip of the iceberg. Here’s how it could transform financial services over the next few years.
Artificial intelligence is undoubtedly the next great breakthrough for the financial services industry.
It will transform the way organisations work, interact with clients and open up all sorts of transformative opportunities. Technology may still be in the early-adopter stage of development, but the next few years will witness change at an unprecedented pace. Riding the wave of this evolution will be profitable, but extremely challenging. To do so, businesses will have to keep abreast of increasingly complex technology, manage cultural transformation and address application challenges.
A report from Gartner suggests that, by 2020, 85% of customer interactions with a business will come from non-human interaction.
Orbis Research, meanwhile, expects the market to grow at more than 50% year on year between 2017 and 2021.
The revolution, then, is coming and it’s coming quickly, and one of the most profoundly affected areas could be banking. Within this industry there’s a race to improve the customer experience, as Keith Bossey, SVP Financial Services at GfK, explains.
“Digital and AI are game changers for financial services because they enable personalisation that humans cannot provide on a cost-effective basis,” said Bossey. “Our data show massive gaps between the benefits consumers want from banks and what they actually get; in many cases, it is only the perceived difficulty of switching banks that is keeping customers in place. The firms that will succeed with AI will find ways to deliver the benefits of human interaction in digital form – a challenge that demands really deep understanding of consumers’ experiences and attitudes.”
How does this manifest itself in practice? Chatbots can deliver enhanced levels of human interaction without the associated cost. Capital One, for example, allows customers to ask basic questions about their bank accounts and credit cards through Skill for Echo. As systems become more sophisticated they are able to better mimic human speech patterns and answer questions more comprehensively.
The importance of data
At the more complex end of the spectrum, banks are looking to harvest the benefits of cognitive computing and machine learning. The development of increasingly complex algorithms will allow machines to mimic the human brain in how it processes information and takes action. To do that, it needs one thing more than anything else: data and lots of it.
Just think about all the data you use to make any decision: You might be consuming information via the rewritten word, processing audio or visual prompts, communications or many more. Computers need to process this all, and turn it into actionable results.
This can help with basic speech recognition patterns but can also enable machines to develop a much deeper understanding of the most complex systems. One of the most important areas could be in fraud detection. By using machine learning banks can process multiple tasks in seconds which would have taken human operatives hours.
Lloyds Bank, for example, has partnered with a US Start-up called Pindrop to detect fraudulent phone calls. The system uses 147 different voice features which can create an audio fingerprint. Representatives are given a traffic light system to alert them if a call is fraudulent. That is then passed onto a fraud specialist.
The machine has the ability to learn and adapt. They let it know if it got the call right or wrong which helps it improve its success rate. At the start of the trial, Pindrop estimated a success rate of 85%. However, as the system evolved and began learning, that rose to 95% during the trial period with Lloyds.
Banks are in an AI arms race with fraudsters.
Cyber criminals are developing their own algorithms to constantly probe a defence for weakness. In response, banks are using AI to assess vast quantities of data to ensure defences are as strong as they can be and identify an attack much more quickly. Indeed, AI will plug the shortfall in cyber security professionals – 1.8 million by 2022.
AI is also being extended to help assess trading risk, provide more accurate financial advice and improve back-office automation. Systems such as GeoQuant are helping banks improve their decision-making by predicting market movements and identifying risks and opportunities. The system, which comes from a US start-up, was able to identify increasing political stability in Italy and Mexico which led to a rise in the value of their markets. Equally, it was able to provide an early warning of increased instability in Brazil which hit stock markets.
The more sophisticated and realistic AI appears, the more data it consumes. Collecting, processing and analysing such data is an enormous technical undertaking for businesses and will involve an enormous upheaval of infrastructure and systems. Departments will have to work together, share information and make data much more agile. After all, the algorithms of the future will be highly hungry for data.
Aside from the technical aspects, culture and working practices are constant obstacles. People and businesses find it all too easy to slip back into a ‘business as usual’ approach. Without the full understanding of the sector, AI can be seen from a narrow viewpoint which underplays its full potential.