Accelerating Innovation With Deep Learning In Financial Services

Technology – specifically that cutting-edge domain of technology that we call artificial intelligence (AI) - is reshaping every sector of business, but few more so than financial services.

In financial services, data and analytics have always primarily been used to crunch numbers, with spreadsheets and business intelligence (BI) platforms long being core tools of the trade. Today, deep learning capabilities – able to analyze and process unstructured data as well as the neatly arranged rows and columns of traditional financial data – mean insights can be pulled from a great many more sources. This means innovative financial services institutions are able to gain a far more in-depth understanding of the market, its customers, and their own internal operations.

However, challenges emerge due to the fact that the market is increasingly fragmented – with agile fintech start-ups, established incumbents, tech hyperscalers, and big retailers like Walmart, Carrefour and Tesco all leveraging their strengths to compete for customers.

Technology as key differentiator

One example of a customer group that is currently targeted by all of these groups for app-based banking and money transfer services are the 66% of the world’s 1.7 billion “unbanked” citizens – many living in areas with poor traditional banking infrastructure - own mobile phones.

Increasingly, the differentiator between success and failure in this outreach will be technology. Whoever most efficiently uses AI – including machine learning and deep learning – to create products and services that solve customer problems and improve their overall experience will be the winner. ?

Recently I spoke with Greg Pavlik, SVP for data and AI services at Oracle, and Kevin Levitt, global industry business development for financial services at NVIDIA ?about the “next level” digital transformation across the financial services industry.

Explaining why AI, in the form of deep learning, has been so transformational in financial services, Pavlik tells me that deep learning algorithms and frameworks have evolved considerably in recent years, to the point where they can now often carry out tasks that would previously have required human intervention, such as reading and understanding a document, or understanding the meaning of an image: “these more sophisticated [deep learning models] can do quite a bit of work that gets closer and closer to the kinds of interpretive analysis that a human would do on a document or image like that … we’re getting more and more sophisticated with the kinds of problems we can solve.”

How is deep learning used in financial services?

There are several key use cases where AI and deep learning have enabled significant growth across the industry. Firstly, in fraud detection, the ability to monitor huge volumes of transactions in real-time and identify which patterns – of behavior, location, currencies, and activity – indicate fraudulent transactions has been a game-changer.

Additionally, chatbots powered by natural language processing (NLP) technology and able to effectively understand and respond to human speech have revolutionized the delivery of customer services and satisfaction.

Another area is risk management – for example, assessing opportunities to invest, offer insurance services, or approve loans and mortgages. Rather than simply relying on statistical analysis of static, structured data, these decisions can now take into account many other sources of unstructured and fast-moving, real-time data sources.

Finally, in marketing operations, more sophisticated decisions can be made about what products and services to offer to customers. This means that these decisions can be based on what is most likely to fit in with the individual requirements of customers, rather than simply what services a bank or insurer wants to sell, therefore improving the customer experience.

Levitt spoke to me about San Francisco fintech NerdWallet, which uses data-driven decision-making to identify not just what products people are likely to apply for but which ones they are likely to be approved for. This enhances the process of product discovery for the customer. NerdWallet uses AI tools powered by NVIDIA GPUs to reduce the time and expense of training their deep learning algorithms by a factor of 10, Levitt tells me.

The Data Lakehouse

Another factor in the evolution of AI and analytics is the growing maturity of storage architecture. Traditionally, banking and financing BI would leverage a data warehouse model, where data is structured and stored for use in analytical applications such as those mentioned above. More recently, this has evolved towards a paradigm known as the data lake, where data is stored in its raw, unstructured format, ready for innovative uses to be applied. Most recently, a “best-of-both-worlds” approach has been adopted by the most forward-looking institutions, which Oracle and others refer to as the “data lakehouse” model. Pavlik tells me that one Oracle customer – credit reference agency Experian – improved performance of AI and data-driven decision-making by 40% and decreased costs by 60% by moving onto a data lakehouse and Oracle Cloud Infrastructure.

The impact of this new breed of smart, cognitive technology capable of leveraging unstructured data is only going to increase as time goes on. Levitt tells me, "We see continued application of AI across financial services, we have banks that have over 500 projects tied to AI, and we believe that while there are hundreds of use cases today, there will be thousands in the future that will ultimately leverage AI to improve operational efficiency, the customer experience, the financial performance – meaning a win-win for everybody involved.”

Both Levitt and Pavlik highlight that one specific area where we can expect to see increasing adoption of AI is in the environmental and social governance (ESG) space. Here, deep learning technology can be used to optimize the efforts of financial services companies to meet their own environmental or social targets, such as reducing carbon emissions or bridging the gender pay gap. They can also use NLP to quickly digest and extract insights from news reports and earnings analyses to help make decisions about who they should be investing in or partnering with.

Pavlik says, "There's a lot of transformation coming, and it can be intimidating, but … it's not something we want to shy away from – it's an opportunity that can drive extremely fast transformation in terms of results for the business.”

Oracle University has a free training and certification offer through 31 December 2021. Visit the offer specifics here

And if you would like to watch my full conversation with Greg Pavlik, SVP for Data and AI Services at Oracle, and Kevin Levitt, Global industry business development for Financial Services at NVIDIA, have a look here.

Thank you for reading my post.?Here?at LinkedIn?and at?Forbes?I regularly write about management and technology trends. To read my future?posts simply?join my network here?or click 'Follow'. Also feel free to connect with me via?Twitter,??Facebook,?Instagram,?Slideshare?or?YouTube.

About Bernard Marr

Bernard Marr is a world-renowned futurist, influencer and thought leader in the field of business and technology. He is the author of 20 best-selling books, writes a regular column for Forbes and advises and coaches many of the world’s best-known organisations. He has over 2 million social media followers, over 1 million newsletter subscribers and was ranked by LinkedIn as one of the top 5 business influencers in the world and the No 1 influencer in the UK.