Friday, September 8, 2023
In this interview, Dr. Mamdouh Refaat, chief data scientist and senior vice president product management at Altair, shared additional insights on the survey and the driving forces behind this emerging technology trend.
The BFSI sector is a leading adopter of digital twin technology. What contributing factors are driving this adoption?
Financial institutions are experiencing more pressure today than they have ever faced before. Emerging competitors and disruptive technologies are changing how and where people spend, save, and invest, with customers demanding more value and flexibility. Criminal activities have become so sophisticated that it’s hard to stay ahead of fraud and cyberattacks. Additionally, regulatory compliance necessitates ongoing self-analysis and reporting. The bar to deliver exceptional products and services has never been higher.
Although the term “digital twins” may be new to the sector, the banking, financial services, and insurance industries have been working with this technology for more than three decades and the process of developing, testing, and deploying digital twins is deeply ingrained in BFSI.
What impacts is digital twin technology having on traditional banking, financial services and insurance practices?
While digital twins are used for a multitude of applications in the BFSI sector, the majority are developed to monitor and predict customer behaviors and address key portfolio performance indicators like credit worthiness, assets under management (AUM), customer loyalty, and customer lifetime value.
For example, banks regularly monitor the health of their credit card portfolios, which is often measured by the overall default rate. This is the percentage of outstanding loans due to regular missed payments (typically 90 days or more). This default rate depends on the profile of the account holders and other general economic indicators such as interest rates, inflation, unemployment rates, and/or global events, like COVID-19.
Leveraging artificial intelligence (AI) and data analytics, the bank can create a machine learning model (a digital twin) that simulates the effects of these different variables on the default rate of a credit card portfolio. The digital twin can then be used to run what-if-analysis simulations to assist in setting up new portfolio management and customer management strategies.
What data points from the survey stood out or surprised you and why?
Although digital twin technology is most commonly associated with the design of physical products in the automotive and aerospace sectors, the survey data revealed that the BFSI sector is using this very same technology to address business challenges like operational efficiencies, security, and fraud detection.
This was no surprise because the BFSI sector has been using the technology for more than three decades (under different names). That was confirmed by the result of this survey that BFSI respondents were the most likely of any industry to say digital twin technology had the greatest positive impact on the "personalization of their products and services." This underscores the critical role digital twins play in delivering exceptional financial products and services to their customers to keep them coming back, given today’s landscape of abundant choices.
What advice would you give enterprises contemplating digital twin adoption
The data gleaned from using digital twins gives organizations new insights that can help finance teams work faster, create better products, generate less waste, and find their next big breakthrough.
Proper implementation of digital twin technology needs three things. First, they need data. Therefore, robust data collection and curation practices are key to any implementation of digital twins. Second, they require know-how and expertise in their techniques of development and implementation. Finally, they need to be implemented to the appropriately defined problems where there is good understanding of the business and its data. Therefore, enterprises who will newly adopt digital twins should start with the most tried and proven areas of applications. This will enable them to benefit from well-established industry-wide experience and best practices in implementing the technology. Areas such as credit scoring, anomaly detection, security, and monitoring investment and credit portfolios would normally be on top of the list.
What should enterprises look for in a prospective digital twin technology service provider?
Digital twin implementation requires collaboration between IT, data analytics, and business units. Partnering with the right digital twin solutions provider is crucial to help foster open communication and a strong understanding of the different aspects required for successful deployment. This includes everything from developing consistent modeling methods, engaging the complete product value chain, guaranteeing access to trusted and accurate data sources, and ensuring continuous progression of the models developed. Given the sensitive nature of BFSI data, the prioritization of robust security measures is also a necessity to protect customer information and financial assets.
Experienced technology providers, with the AI and data analytics know-how and capabilities, can help enterprises think through the many structural and organizational changes that need to happen within the organization as well as the technology infrastructure, software, and data analytics tools to enable the deployment of digital twins. With the recent advances in Large Language Models (LLMs), digital twins will benefit from integrating this new technology in BFSI specific applications.
For example, in most cases an army of data scientists are needed to develop a BFSI ML model, including the validation and testing of the model, as well as monitoring its performance. But with the application of self-service data analytics tools, organizations don’t have to hire data scientists (that are already in limited supply). Instead, the same finance professionals who are experiencing the day-to-day challenges and problems their business and/or customers are facing are the ones developing the digital twin models to influence better outcomes and decision making.
How has digital twin technology enhanced fraud detection and prevention?
By combining digital twins with LLMs, BFSI institutions can analyze large amounts of data in real-time and automatically flag suspicious behaviors that are commonly associated with fraud. Additionally, with the ability to run simulations using different variables (potential risks), teams can also identify risk factors and areas of their operations most susceptible to fraud while eliminating the need for manual prevention.
What near-term trends do you expect to see for digital twin technology?
Digital twin and Machine Learning technology providers now offer products and services that speed up the cycle of implementation. Many providers' current offerings not only allow data scientists to develop digital twins, but also enable domain experts and business analysts to develop them and use them alongside the data scientists.
This was a result of strong investments in the field, as well as, applying AI technology to automate many of the steps that data scientists typically take. This new situation will offer many organizations an easier entry point to developing and implementing digital twins. It will also free data scientists’ time to develop solutions to more complex problems for novel application domains.
Additionally, software providers may begin offering out-of-the-box digital twin solutions, for specific business applications, that may require little tuning to be implemented in the field. Some providers are developing their “marketplace” platform to offer such solutions. 
The Green Sheet Inc. is now a proud affiliate of Bankcard Life, a premier community that provides industry-leading training and resources for payment professionals. Click here for more information.
Notice to readers: These are archived articles. Contact names or information may be out of date. We regret any inconvenience.
