Financial Services: Taking AI To the Next Level
Parts of the financial sector, notably wealth management and equity trading, have been among the first movers in the commercial use of AI. The scope of AI-led innovation is now widening: banks and insurers are actively applying AI techniques in the front and back office, developing innovative customer-facing services and automating operations such as payments, risk modelling and fraud detection.
One of the key drivers for such swift adoption of AI in the financial industry is cost savings. By 2030, traditional finance sector organisations could reduce their costs by 22%, according to fintech research company Autonomous Research, in what would amount to more than US$1 trillion in efficiencies. These savings would come from the front, middle, and back office operations and includes a reduction in retail branches and bank tellers, the application of AI to compliance and data processing, and the automation of underwriting and collections systems.
Outside of the financial motivations, technology companies—notably small fintechs— have provided the impetus for much of the innovation itself. Some established financial sector players have responded energetically themselves, co-opting fintechs’ AI expertise and innovations. But Andrei Kirilenko of Imperial College London maintains that the variety of services traditional banks and insurers offered remains limited. He expects a new wave of innovation as technology companies exert more pressure on the industry to meet customer demand for smarter, more personalized products. “New financial products that no one heard of before will proliferate,” he says.
More Data, Better predictions, Faster Services
What directions will AI-led innovation take? One will be a marked improvement in capabilities that exist today, as machine learning tools analyze ever larger amounts of data in ever wider varieties. For example, equity trading algorithms used by investment management firms and hedge funds should gain in predictive accuracy, something which has considerable room for improvement, according to some analysts.
Market Strategy and Potential Targets: Example, AI Robo Advisers for Both Emerging and Mature Markets
According to William Genovese, Vice President of Corporate Strategy Planning, Banking and Financial Markets, in terms of one very popular use case – robo advisers for investments, “I believe that target markets should be centered around two customer segments to ‘optimally get the job done’ 1. Use robo advisers at the lower end while marketing it to new investors who would welcome robo advisory as a self-service tool for its simplicity and flexibility. In most emerging markets, the rising number of middle class and upper middle class are looking for opportunities to invest; however, the lack of knowledge and access to information prevent them from investing. For the lower end disruption of the market, simplicity is the defining characteristic. The design of robo advisers should initially be simple and easy to understand by the investors. Basic applications that are publicly available in the market like Bloom, Robinhood, Stash, and Acorns serve the investors’ need and demand in the lower end market. For instance, Stash offers fund investments based on lifestyle preferences and interests (that is, clean water technology companies, Internet companies, and so on). These applications are straightforward and simple to use and readily adaptable and consumable by these new customers. Due to these features, investors can quickly learn how to invest – they can test the markets based on their life interests and causes by investing in a collection of companies that reflect their desires and goal accordingly.”
To thrive in such a landscape and to flourish in both mature and emerging markets, established players will need to embrace more active partnerships with fintechs, says Huawei’s Genovese. In banking, for example, innovation will likely to increasingly originate in platform-based ecosystems, rather than within banks themselves. “Traditional players,” he says, “need to embrace such forms of collaboration to compete effectively. Effective and secure structured and unstructured data management for accelerated insights is what these platform based ecosystems have now, and is the solid foundation that AI needs to provide the most value. 50% of all banking customers will give their financial institutions two chances to get the customer experience “right” with them. If they don’t get it right, the customer is going elsewhere, and the customer usually is based on how convenient financial services is embedded in their lives. AI is harnessing this customer data through integrated platforms very quickly to make this happen”.
The chatbots that banks are now using for customer interaction will similarly use wider data access and continuous learning to make more targeted offers to customers and provide more accurate remedies to resolve issues. In five years, such virtual assistants may rival websites and the physical branch in importance as banking channels for customers. In a global survey conducted by The Economist Intelligence Unit in early 2018, banking executives pointed to “improvement of the customer experience” and “greater customer engagement” as the main benefits their firms will generate from AI use in the next five years.
Banks and non-bank lenders will use AI to accelerate loan approvals, which are already being automated today. Real time lending decisions will become widespread, as credit scoring and risk appraisals are similarly automated.
AI will also come into use in the back office, as banks push ahead with the automation of payments, fraud detection, risk management, compliance and other operations. Much of this is currently driven by robotic process automation (RPA), but predictive analytics and machine learning will begin to play a larger role in these areas within the next couple of years.
Cross-sector Germination
New payment models and services are likely to emerge as financial service providers gain wider access to data from other sectors. The insurance industry’s access to automotive data, generated by in-car IoT sensors, is leading to the growth of usage-based car insurance, for example, in which AI-based assessments of driver behavior factor into “pay-how-you-drive” premiums.
Dr Kirilenko similarly expects the ability of insurers and banks to apply AI techniques to the analysis of customers’ health data—generated, for example, from personal health monitoring devices—to result in new types of financial products. Wealth managers will also craft investment products and advice to fit customers’ health profiles. Stricter data privacy rules in some markets may impede such sharing of such data in the short term, but cross-sector data flows will inevitably increase, he believes, as customers come to expect more joined-up financial, health and other lifestyle products.
Disruption to Come
There will also be change, believes Dr Kirilenko, in the mix of players able to offer financial services. Combining their AI and broader digital expertise with their unique ability to glean customer behavior and preferences, large technology companies will come to compete more directly with traditional financial institutions. “Regulators in many markets are more open today to the idea of replacing parts of the financial systems with technologies that work,” he says.
Action Required
The uses of AI detailed above require that firms have access to ever greater amounts of computing capacity to support the more powerful algorithms they will run. Banks and insurers have been less avid users of public cloud services than firms in other industries, due partly to perceived risks to customer and proprietary data. That is likely to change as cloud service providers assuage security concerns, and as banks and insurers become more active partners with fintechs and other organizations —including even competitors — in innovation-oriented ecosystems. Many companies will find that they cannot take their AI initiatives to the desired scale without greater use of the cloud.
The admonition that organisations’ data assets must be readied for AI uses applies to all industries, but the challenge may be greater for established financial industry players. For example, research conducted by the consultancy PwC in early 2018 found that financial services firms were among the least effective in their use of data in a cross-sector comparison. Only 26% of financial sector executives said their firms use data effectively, third lowest of ten industries covered. There is much to do to redress this, but industry firms will need to integrate as many siloed data sets and systems as regulation allows, as to be effective AI needs access to wide varieties of data, and in different formats. Banks and insurers both will also need to deploy analytics tools that are adept at working with unstructured data.
This is not entirely uncharted territory for established financial industry players, and many are moving ahead aggressively to build AI capabilities. But if the borders separating the financial and other sectors erode as Dr Kirilenko predicts, many financial institutions will find it difficult to keep pace.
Editor’s note: This article is a guest post from BrandConnect
Produced by (E) BrandConnect, a commercial division of The Economist Group, which operates separately from the editorial staff of The Economist and The Economist Intelligence Unit. Neither (E) BrandConnect nor its affiliates accept any responsibility or liability for reliance by any party on this content.
Lead contributor: William Genovese, President of Corporate Strategy, Research and Planning – Banking, Financial Markets and IT Services at Huawei
Disclaimer: Any views and/or opinions expressed in this post by individual authors or contributors are their personal views and/or opinions and do not necessarily reflect the views and/or opinions of Huawei Technologies.
Nice knowledge-gaining blog. This post is the best on this valuable topic. I like your explanation of the topic and your ability to do work. I found your post very interesting