Home > What’s UP > blog > AI in banking: know your basics

blog

07. 06. 2018

AI in banking: know your basics

How to use AI to cut costs and grow revenues? This is one of the most talked-about questions in banking today. But not all bankers have a grasp of what the new technology can do, let alone what it can do for them. Let’s see how the AI magic works and why one sector after another is falling under its spell.

Retail, manufacturing, telecommunications: these are some of the industries, where AI already plays a crucial role in boosting productivity, reducing costs and providing in-depth customer insights. Not to mention the technology sector, where big players like Amazon or Alibaba are making AI part of their game plan in various shapes and forms. The growing competitive threat from tech behemoths and digital-only banks should be a wake-up call for all incumbent financial institutions that it’s time they dipped their toe into AI as well.

Many of the leading banks already have. And they’ve achieved considerable success in applying AI-driven tools, such as chatbots or virtual assistants, to improve user experience and increase customer engagement. Today front office AI applications dominate the financial services sector, the Financial Times has found in its recent survey of 30 of the world’s biggest banks. Seventeen of the 18 banks that had provided detailed answers are already using AI in their front office operations, ranging from Citi’s Facebook messenger chatbot to UBS’s use of Amazon’s virtual assistant, Alexa, in customer service.

AI in banking sales: a $200 billion game

But digital banking sales seems to be lagging behind when it comes to putting AI to work. Lenders who are not looking into using AI risk a lot: AI-powered applications could unlock a potential value of a massive $200 billion in banking sales and marketing alone, according to McKinsey. In particular, artificial intelligence and machine learning have significant value potential in retail banking, much as it does in retail. Machine learning, a subset of AI primarily used to provide predictions and prescriptions, is one of the most important AI tools for banks in digital sales.

While consumer-facing AI tools can improve customer service, internal uses of AI make it possible for financial institutions to create far more personalized products than ever before. Adam Shardlow, Lead Journey Manager for AI at the Royal Bank of Scotland has told W.UP in a recent interview that the traditional banking model had always been around ‘one size fits all’, but that model had required that customers have a full-time job, own a property and a car, and have two or three children. Going forward, that is not the way customers will necessarily live their lives, and banks will use AI to create products in a far more personalized way, he reckons.

How AI works: not very easy like ABC?

This all sounds great? It does. But let’s cover the basics first. So what does artificial intelligence mean after all? In a nutshell, AI refers to technologies that perform tasks normally requiring human intelligence. These tools can carry out a wide variety of functions, from recognizing and understanding human input to learning independently from data patterns and applying context to interact with customers, Deloitte summarizes.

AI does not represent a single technology. It embodies a variety of technologies, methods and capabilities that often support each other in a multidimensional field. The multiplier effect of these technologies, powered by advances in cloud computing and processing capacity, has also multiplied the significance of AI in several industries, including banking.

The use of AI can be basically grouped into three domains: automation, engagement and insights. In the cognitive automation domain, robotics process automation (RPA) and other tools are used to develop specific expertise and automate jobs that are traditionally performed by highly trained human workers.

Engagement, the second domain, includes AI systems employing cognitive technology to connect with customers. In the third domain, AI technologies, like machine learning, are applied to create actionable insights by extracting concepts from data streams, and generate personalized and relevant answers hidden in masses of unstructured data.

Deep neural nets in ML are able to harness data, which a lot of the more traditional banks have had real problems utilizing, according to Adam Shardlow. These banks have traditionally built in silos, but in the future, all these data will be brought together, and institutions will have a single view of the customers and be able to create far more predictive solutions for them.

The banks that benefit most from AI will be those that are prepared to rethink their approach to their people, their processes and their data, according to Accenture. Artificial intelligence is expected to break through the silos and practices of process-driven banking, allowing banks to become analytics-driven entities that use data to dynamically inform and shape what they do in real time.

The future of AI: hopes, hype and the reality

Financial institutions have high hopes for AI. Seventy-one percent of bankers believe that AI is capable of becoming the face of their organization or brand, and 40% plan to invest in embedded AI solutions in the next three years, according to a survey by Accenture. Swiss lender UBS has found in its survey of 86 banks that AI technology could potentially lead to a 3.4% revenue uplift and cost savings of 3.9% over the next three years.

But some experts say the current hype around using artificial intelligence in banking needs to be “tinged with an air of caution”. Why? Infosys Finacle says the technologies supporting AI solutions are probably not quite as ready for real-time use as the excitement would suggest. And one of the hurdles banks may face “as they delve deeper into the AI world that promised so much” are unrealistic expectations, the Financial Times has warned. Too much investment may flow into sexy areas, like chatbots, at the expense of investment in internal processes potentially offering bigger gains.

Whether or not these possible risks and dangers are real, is yet to be seen. In any case, there is still room for banks to grow in AI. The financial services sector is clearly not very mature in adopting these solutions, as it came in at the eighth place in Infosys Finacle’s rank of various industries regarding their AI maturity index. Pharmaceuticals and life sciences, automotive and aerospace, and telecoms had the highest maturity scores in the list.

AI and ML in digital banking sales