Digitisation lags in corporate banking and insurance

Every bank CEO is keen to tell us how they are technology companies today. They emphasize their thousands of software engineers and APIs. And why not. Until recently the technology sector was booming. Technology is foundational to innovation in all industries. Financial services are no different. If used correctly technology drives competitive advantage, scale economies and higher return on equity. But the use of technology varies massively across financial services. I started my career in banking on the trading floor of Credit Suisse First Boston (as it was known) in the late nineties. I still remember the noise of walking on a trading floor in those days with hundreds of traders, sales traders and salespeople shouting prices down the phone all day. Today, the financial markets are digitised. Most trading is electronic and manual traders are increasingly replaced by an army of quants and technologists. We are also seeing huge digital transformations in the payments business and retail banking user journeys. By contrast, other parts of financial services such as corporate lending and insurance are lagging in their adoption of technology. This, even though insurance is the original data-driven business model. Humans will always be crucial as risk managers to assess the credit worthiness of borrowers and in the underwriting and pricing of insurance. But the amount of data available to corporate bankers and insurers has boomed in the internet age. And if incumbents don’t use this data other players whether it be nonbank lenders like private equity and hedge funds or InsurTech firms will use this data to create a competitive advantage. Algorithms whether they be rules-based or AI-driven can help in digitising previously manual processes in corporate banking and insurance. At Galytix, we offer an end-to-end data ecosystem from data discovery and ingestion through to transformation and analytics. We digitise previously manual processes saving credit analysts 40% of their time and increase their time to market by 30%. This allows them to focus on refining their models and new requirements such as ESG. Our data ecosystem combines structured and unstructured data, financial statements from annual reports with news flow, and internal and external data in a systematic fashion. Banks need one connected data ecosystem that brings together all data on corporate clients rather than the traditional siloed approach. This is crucial for risk managers that are dealing with the twin challenges of rising credit defaults and increased KYC requirements stemming from the Russia sanctions. Many large corporate banks and insurers still use a small number of data sources and rely excessively on internal data directly from their clients and counterparties. Our neural network discovers data from more than 240 diverse sources at speed and scale. This includes filings of annual and quarterly accounts by public and private companies, court and bankruptcy filings, industry data available across a wide range of regulator and government websites, equity and fixed income analyst ratings and estimates, management changes and insider selling and a filtered approach to news flow around other crucial risk factors like fraud, cyber and climate risks. World-class AI algorithms are a solid foundation, but we ensure high levels of data and model accuracy and reliability by taking a software plus services approach. AI is great at giving the speed and scale that humans physically cannot achieve in gathering data from such a wide range of sources and processing this data. At the same time, our sophisticated large bank and insurance clients expect human-like data accuracy. When the structure of a new data type or document is completely different to what has been processed previously, the algorithms may need re-training and human supervision is needed to guarantee 100% accuracy. Similarly, back-testing on historical data may not always be the most predictive of the future in a constantly changing world and human supervision by the data engineering team is needed to fine-tune these algorithms. Technology alone is not a magic bullet. Banks and insurers that will win have business leaders who know what data is relevant, how this data needs to be classified to generate useful insights, and how this usage should be governed. The technology function can’t be an island where developers are not focused on business outcomes. Both business and technology leaders in banks and insurers need to understand what parts of the value chain need algorithms and which parts need human supervision. Given the complex and rapidly changing environment, algorithmic models will need to be combined with expert judgments.