Mastering AI and Big Data Analytics: Why Fintechs Must First Get Their Data in Order
The global fintech industry is continuing to thrive, and its desire to embrace artificial intelligence (AI) and big data analytics remains paramount.
Ambitious and dynamic, the 20,000 or so fintechs currently operating globally know AI, machine learning (ML) and big data solutions will be the keystones for success or failure in the near future. These technologies will enable them to differentiate their offerings in a market where competition is intense and the pace of innovation is fast and unrelenting. In effect, the adoption of AI, ML and big data analytics is the price of admission to the main event in banking and finance because potential partners or customers want fintechs that stand out and deliver much more than the status quo.
The mainstream finance sector is already employing AI and Big Data solutions, where the best-known deployments cover algorithmic trading, market monitoring, and fraud detection. In algorithmic trading, big data analytics enables a flexibility and sophistication that no human can match. The major market names use this technology to execute large trades throughout the day without triggering adverse price movements. While financial data analytics enables risk-based pricing, AI can spot suspicious patterns that indicate fraud or money laundering activity.
However, despite the advances in AI and ML, the banking and finance sector is still held back from the adoption of more complex models by interoperability problems, legacy systems and a lack of qualified talent. There is also the question of regulation, particularly in the areas of data governance, risk, traceability, auditability and the elimination of in-built bias.
For the fintech industry, these difficulties faced by banks and financial organizations present multiple opportunities. AI will enable fintechs to offer very high levels of consumer personalization and to provide financial services that are faster and less costly than the established institutions – yet do not require a vast array of expensive back-office systems. Getting this right is a potentially huge advantage for fintechs, but it demands the most effective possible handling of financial data.
Make data quality and management the twin priorities
With this clear backdrop of varying challenges, how should the fintech industry and financial services organizations implement AI and Big Data analytics and what should be their priorities? The answer must be to sort out the data first.
Despite their innovation, one of the most substantial challenges for fintechs remains data quality. Without fast and easy access to the right kind of data, the deployment of AI and analytics will fail to deliver the transformation of which these technologies are capable.
To fully exploit AI and analytics, unstructured data must be merged with structured internal data. SQL databases are examples of structured databases, covering transaction information and credit systems. These have been around for a long time. Unstructured data includes log files, videos, emails, and social posts. For the banking sector, this is unfamiliar territory, and its integration requires care and expertise. No surprise then, that where institutions have begun using AI, they primarily employ their own internal data, rather than data from external sources such as social media.
To fully leverage AI and big data analytics using all the data available and not just the internal sources, fintechs must address quality. This is to ensure accuracy and trust in the data, with a clear understanding of its sources and lineage. These questions must be resolved before organizations start building or deploying applications around descriptive, diagnostic, predictive, and prescriptive analytics. Prescriptive analytics recommend a specific action be taken and will open the path to highly sophisticated automation.
Financial data volumes are expanding rapidly
As well as addressing data quality, financial services organizations and fintechs must overcome the challenges of data quantity. This is vital because the best and most accurate analytics often require masses of data. The good news is that there is plenty of it in the wider world, where we have reached 59 zettabytes (ZB) of data and are on course to hit 150 ZB in five years’ time. The difficulty is in handling the volume – to use and integrate it to provide the results the organization needs.
Agility is also essential. Organizations must be able to excel when data volumes suddenly spike caused by real-time market changes or surges in transactions that generate increased levels of streaming. Such spikes are part of life in the financial world and must be accommodated.
Success or failure depends on the approach to financial data analytics
For fintechs, as much as the financial institutions they serve, success with AI and big data analytics therefore demands a more integrated approach to data, capable of handling the myriad data management requirements.
Adoption of a smart fabric is the most effective approach to this primary challenge, interweaving all kinds of information from many sources, transforming and harmonizing it for the needs of highly data-intensive applications. The smart data fabric approach enables an organization to excel with fast-moving, batch-oriented data, along with data from cloud and legacy sources, connecting where required through APIs or web services.
The smart data fabric also possesses the critical ability to scale dynamically to cope with surges in data volumes, making information usable through a simplified, streamlined architecture. This is a location-agnostic approach compatible with multiple cloud storage systems, which provides consistent access through micro data services.
The alternative is to struggle in a tangle of unmanageable, dubious data that before long, nobody within the organization can trust. If they are built on poor quality data, AI and big data analytics will be incapable of increasing competitiveness, profitability or higher levels of service and personalization.
Badly managed data may even put an entire organization in danger through operational, financial, and regulatory failure. Bad data also makes it hard for any fintech organization to change, and without change few businesses will survive in a field of such intense competition. By turning information into an asset, data becomes the rocket fuel for dynamic change and growth, enabling fintechs and their partners to excel in the era of AI and big data analytics.
Find out more about the future of fintech and why a data-driven approach will be key by playing back our webinar with Fintech Ireland: http://bit.ly/38XK8c8
Michael Hom is a technology executive with over 25 years of experience working in Financial Services industry. Prior to joining InterSystems, Michael was a Managing Director managing Global Rates, Securitized Products, and Municipals Technologies at Royal Bank of Canada Capital Markets. Previously, he was an Executive Director overseeing Cross Product Technologies including Risk, Sales and Trade Management at Nomura Securities. Michael started his career at Lehman Brothers, building systems in the Rates, Foreign Exchange and Emerging Markets areas. Later on, he became Senior Vice President leading Securitized Products – Whole Loans, Real Estate, and Principal Finance Technologies. He holds a Bachelor’s degree in computer science from Columbia University School of Engineering and Applied Science.