In search of the Amazon of Fintech –
Full disclosure – before working at Analytics Ventures my impression of Artificial Intelligence was almost entirely driven by science fiction. The drive towards general AI inevitably led to fears of out of control robots destroying everything I’d held dear, or a kinder dystopia where Scarlett Johannsen would whisper gently into my ear as I fell asleep or cooked breakfast. In terms of applying this to the world of finance, my experience influenced my thinking.
Having worked for a Hedge Fund for over four years in the UK I had a huge amount of respect for the analysts that could balance underlying idiosyncratic fundamentals with competing disparate economic and political factors. Despite this, the concept of automating their role entirely didn’t seem too farfetched due to the ability of algorithms to quickly process large data sets far beyond the capabilities of a human being. As a result, when I started at Analytics Ventures, I assumed that this would be the first major application of AI in the financial industry. However, the more I’ve gathered about the real-world applications of machine learning (ML) and the opportunities for optimization across business lines and industries, the more the current state of AI in the world of finance makes logical sense.
At present, large financial institutions are hesitant about spending reporting cycles and significant expenditure on capital-intensive attempts to fundamentally disrupt the industry. Whilst, according to Forbes, 65% of senior financial executives expect to see the positive impact of AI, as of late 2018 only a third of financial institutions claim to have successfully integrated AI/ML solutions into their operations. Part of the reason for this hesitancy is that, at present, data quality is impeding their ability to deploy effective solutions. A phrase that is pertinent here (and one I’ve heard several times in Dynam.AI meetings with potential clients) is ‘garbage in, garbage out’. Unstructured data, along with data collected from a wide range of sources needs considerable amounts of work before insights are truly reliable. Alongside this, the ability to hire the talent required to quickly extract usable insights from this data is in short supply. These two factors are fleshed out in an April 2019 Refinitiv survey of 450 financial professionals. Of those surveyed, 43% say poor data quality is the biggest ML barrier, 38% report a lack of data availability as the 2nd biggest challenge, and 75% find it hard to hire the required talent.
As a result, the early adopters are increasingly focused on ‘quick wins’- the exercise of rapidly implementing AI projects that have an immediate, visible ROI. These projects allow AI decision makers to display to leadership how valuable AI technology can be and to generate excitement internally about a future AI-driven strategy. Accordingly, the AI initiatives we read about being implemented by financial institutions are often very similar to those we read about in other industries.
Robotic Process Automation (RPA): ML allows firms like JP Morgan Chase to verify data, generate reports, review documents, and extract information from forms. This makes it possible to automate a number of time-consuming tasks; cutting costs and boosting productivity. Ernst & Young recently reported a 50-70% cost reduction when RPA is applied, and Forbes called it a ‘Gateway Drug to Digital Transformation’.
Increased knowledge of customers: smart chatbots and virtual assistants leveraging Natural Language Processing (NLP) are increasingly used to provide customers with real-time assistance while reducing the labor costs associated with call centers. Alongside this, firms are leveraging ML technology to improve and personalize the customer experience, offering financial guidance through the prediction of spending habits and behavior. Bank of America, JP Morgan Chase and Wells Fargo have all launched banking apps that take advantage of these technologies.
Fraud Detection: ML has also proven itself unique in its ability to combat financial fraud and money laundering. ML solutions used at institutions like Goldman Sachs, American Express, and Citibank are able to detect anomalies in customer behavior, buying habits and location. This, in turn, allows them to quickly recognize and prevent suspicious activity.
Risk management: With the amount of real-time data being generated in the world of finance, ML technology is able to make faster, more accurate lending decisions. This technology also allows firms to adjust forecasts based on new risk factors beyond the processing speed of traditional rules-based systems.
Mergers & Acquisitions: Similar to the growth strategies of larger tech firms, the appeal of acquiring fintech start-ups to acquire new customers and accumulate expertise has been one of the defining factors in the adoption of AI.
As you can see, AI is finding its foothold by addressing a number of challenges in the financial sector that have existed since the creation of modern banking. However, as yet AI has not found its own industry-specific gamechanger in finance akin to the way that Amazon was able to harness the power of the internet and personal data to fundamentally disrupt the world of retail. Your turn, AlphaTrAI.
About the Author
Jack Boath is generating new clients and sourcing funding opportunities for Analytics Ventures, a San Diego based venture studio with Artificial Intelligence at its core. Jack brings a strong background in Investment Management and Investor Relations, having worked for a global hedge fund with equity, credit, and structured credit focus. Born and raised in the UK, with an MBA from the University of San Diego and a degree in economics from the University of Manchester, Jack is passionate about football (soccer for you Americans), music, and conversation.