Making big leaps in Banking with Big Data

The banking risk management scenario is steadily evolving. The system of compiling data from information silos and feeding them into manual spreadsheets is now fading into the past.

Ramkumar Iyer
Lead Technical Architect at BCT Digital

Rajiv Singh
Principal Consultant at BCT Digital

Making big leaps in Banking with Big Data

The banking risk management scenario is steadily evolving. The system of compiling data from information silos and feeding them into manual spreadsheets is now fading into the past.

Ramkumar Iyer
Lead Technical Architect at BCT Digital

Rajiv Singh
Principal Consultant at BCT Digital

The banking risk management scenario is steadily evolving. The system of compiling data from information silos and feeding them into manual spreadsheets is now fading into the past. Claiming its place is a dynamic ecosystem teeming with petabytes of information, collated in real-time from hyper-connected networks and sources. And at the epicenter of this set-up is big data — forever interpreting, analyzing and enabling transformation.

Recent banking reforms are forcing the industry to relinquish age-old techniques of data analysis and modelling, and adopt more rewarding technologies. With digitization, the volume of data has been increasing in leaps and bounds. It is becoming evident that the technology available with most banks is not adequate enough to process all of this data being generated at an unprecedented pace. Not only is an alarming amount of data going to waste, along with it, valuable insights are being lost to the industry.

Enabling the paradigm shift

Luckily, technology advances have facilitated the creation of tools that could handle previously unimaginable amounts of data. There is no longer a dependence on data that is “clean” and computationally manageable. Thus, banking industry is adopting such new and emerging technologies — like big data, AI/ML and analytics

  • Making sense of unstructured data: All around us, massive volumes of data are being generated every day. Most of this is unstructured data — an assortment of different data formats, structures and types. Traditional risk management practices that rely heavily on structured data are rendered inefficient. Using advanced technologies like big data, data analysts can make order out of the chaos and extract critical insights.
  • Single source of truth: The influx of data from various touch points — the news, social media and so on — begs the question of data duplication. Big data products allow large volumes of data to be stored at a single repository, open to query and analysis by different departments within the same organization.
  • Customer-centricity: Traditional banking practices relied on small data sets that were customized to specific data models. Big data can go beyond the restrictions of such information silos, pulling data from multiple channels. This enables layered perspectives and 360-degree analytics, and is immensely valuable in areas like customer profiling for risk mitigation.

Rise of the private banking sector

When it comes to big data adoption, emerging economies like India are yet to pick up pace. The current regulatory landscape being unfavourable to data storage, privacy and access are some of the reasons. In India’s public sector banks, for example, big data adoption is still a work in progress.

In stark contrast, private players were quick to catch up, rolling out new products and initiatives at par with even some of the global counterparts:

  • Self Service BI: As we engineer systems that learn from experience, fraud risk management takes on a new perspective. Deep-learning systems can identify patterns in ways that cannot be coded through rule-based algorithms while detecting potential risks faster than any human expert.
  • Integration of big data with Blockchain: As a universal and immutable source of truth, Blockchain is an authority in curbing financial frauds. The verifiable nature of Blockchain data simplifies analysis. When combined with big data analytics, permission-based Blockchain networks could become immensely valuable as the future of risk management.
  • Tackling cyber risks: Digitization has opened the banking system to online threats. Banks, in turn, are eyeing big-data-based security tools to curb sophisticated cyber threats. Big data can be used to flag systemic security gaps, detect intruder attacks, and launch defensive maneuvers and so on — serving as a highly effective threat monitoring and mitigation mechanism.

Indian PSBs: Roadblocks ahead

Despite the obvious advantage of big data and related technologies, few of the PSBs are embracing this advantage, while there are many others in progress. The push from regulators to roll out advanced technologies into everyday banking is also fairly recent.

Additionally, from a governance perspective, banks are working towards imbibing good data management practices. This includes understanding technologies and their application, managing multi-level integrations, and ushering in the transition. Experienced service providers can help banks with the right advisory support, overcome barriers, and foster effective changes in the organizational hierarchies.

Technology partners must take additional effort to help banks with the nuances of new and emerging technologies. They must be candid in detailing the pros and cons of big data adoption, and act as thought leaders in overcoming industry-specific challenges. Such transparency — which is at the core of all professional relationships — will be hugely helpful in promoting innovation.

Big data, big questions

With the glory, come the questions. The advent of big data raises concerns of data security and authenticity. There is a growing fear that technology is being used for surveillance, and manipulated to serve unjustified purposes.

The below troika raises pertinent questions on the industry’s move to big data analytics:

  • Volume: With the impending data explosion, can the volume and variety of data (the curse of dimensionality) become too much for big data systems to handle effectively?
  • Veracity: As most big data and AI algorithms are black boxes, can regulators really determine their fairness?
  • Volatility: Can all analytical data be retained indefinitely, and if it cannot, will this impact the results of the analytics?

How much of this concern is real? Are users today really in control of their data?

As with all technology, the transformative power of big data and AI needs to be taken with a pinch of salt. When used prudently, big data is revolutionary in its ability to solve specialized problems. It can not only analyse data with remarkable speed, but also discover solutions to problems that would take the average human intelligence years to solve. With this superpower of an ally at their disposal, manual efforts can be focused where they matter most — in strategizing and driving growth. Therein lies the power of big data.

Secondly, every country follows certain set standards for security and privacy. The fundamental tenets of data regulation prevent big data and AI algorithms from discriminating (fairness) or revealing identifiable information. Meanwhile, pre-set techniques analyse data without revealing unique information. Stringent regulations like that of EU’s General Data Protection Regulation (GDPR) enforce this practice.

And last, every technology deployment is rooted in the belief that if caution and prudency are compromised, things are bound to go wrong. Right now, banks, including PSBs, are steadily moving towards customer centricity. Proactive customer service with quick turnarounds has become the need of the hour. For PSBs to enter and beat private players at their own game, a paradigm shift is needed in the status quo. And it takes the right technology and the right service provider to make this shift happen.

rt360 — leading the next generation of banking technologies

The next-wave of FinTech applications led by experienced service providers can make a world of difference. One such example is rt360, built by the bankers-turned-solution-architects of BCT Digital — Bahwan CyberTek. It is a powerful product underpinned by a “Business First, Technology Next” approach, that helps manage, monitor and track models throughout their life cycle.

The industry and technology experts at BCT begin every engagement with the “discovery phase”. The as-is system landscape and processes of a bank are assessed and evaluated at this phase to broaden our understanding of the operations. Every custom development is engineered by taking account of global best practices, industry regulations and our own industry expertise. This accelerates the deployment process while reducing errors and re-work, keeping mission-critical processes afloat.

Some of the proprietary technology innovations fuelled by rt360 include:

  • Early Warning Systems to predict credit defaults through advanced algorithms and rules engine-based analytical set-up
  • Text miner tool using Stanford NLP (Natural Language Processing) to extract critical information from high-volume data pools, and issue alerts
  • Internet scanner to report relevant customer information from online data sources using fuzzy logic; serves as an effective tool for data aggregation, asset and fraud/risk management
  • An incremental upgrade architecture along with an aggregator to prevent duplicate data fetching. The architecture involves requesting only the changed entities and areas of interest on a regular basis (using JSON). This optimizes the network and storage times.
  • Mapping of JSON data to relational storage structures done in multiple innovative ways, without using a native JSON database due to commercial considerations
  • Usage of Apache UIMA Ruta annotation scripts to detect Stock Audit Format
  • Network analysis and inter-linkages of cross directorships (to depict the interconnected nature of firms) using graphical databases

rt360 saves users time and cost by continuously updating the product with new compliance mandates that come from the global regulators. It is presently deployed at eight of the top ten banks in India, helping them to meet their risk management and compliance mandates.

Authors

Mr. Ramkumar Iyer

Lead Technical Architect at BCT Digital

Mr. Ram is a full stack Product Architect with around 20 years of experience in software engineering and product development. His areas of expertise spreads across Product Engineering, Solution Architecture, and PreSales. Technologies he has worked on include Java/JEE, SOA/Microservices, Python, AngularJS, ReactJS, API design(SOAP/REST), Cloud (AWS, GCP), Mobility, AI/NLP and Open source software with premier telecom and banking clients in Agile mode.

Mr. Rajiv Singh

Principal Consultant at BCT Digital

Mr. Rajiv comes with 15+ years of professional experience working through Model Development and Validation for banks across the globe. The data scientist in him has also contributed to the development and implementation of Integrated Risk Management Solutions. He is an MBA Finance & M.Sc.(Comp Sc.) graduate and is also certified in a couple of other fields as Financial Risk Manager (FRM), BAI-IIM-B and PGP-AL/ML.

Share On

Also Read

aibased, Thought Leadership
By rt360 January 13, 2020

AI-based Models and Model Risk

AI-based models are more susceptible to certain risks than their conventional counterparts, unless they are put through proper governance and oversight mechanisms. We will look at some of the model risks AI models may be susceptible (over its conventional counterparts).

Kasthuri Rangan Bhaskar
VP, Financial Services Practice & Risk SME (Lead) at BCT Digital

Read More

machinelearning, Thought Leadership
By rt360 December 23, 2019

Machine Learning and Credit Fraud Detection

Machine-Learning-based algorithms lend themselves as useful tools for analysing the possibilities of credit fraud. This is typically done by running the algorithms on various transactional data and analysing for hidden patterns.

Kasthuri Rangan Bhaskar
VP, Financial Services Practice & Risk SME (Lead) at BCT Digital

Read More