One-size-fits-all GenAI solutions may be challenging, while tailored solutions may drain budgets. How can fintech firms strategize in the Goldilocks Zone?
In the banking, financial services and insurance industry (BFSI) generative AI is a major talking point among early adopters as well as those sitting on the fence; and those confused by the hype and clutter around the technology at any point in their consideration, implementation or management.
With increasing scrutiny and regulation over the need for organizations to adopt Responsible AI, as well as the greater exposure to cyber risk if GenAI is not implemented securely, how can the industry tap into the tech with minimal friction, and within the organizational goldilocks zone?
DigiconAsia.net received useful inputs from Dipen Mehta, APAC Head of Banking, Financial Services, and Insurance (BFSI), Softserve to share with readers…
DigiconAsia: What do you think are the factors compelling certain business leaders to rush into adopting (or not rushing to adopt) GenAI?
Dipen Mehta (DM): The emergence of GenAI is disrupting the BFSI sector by enabling faster innovation and improving operating models in some areas.
We see GenAI being probably the first piece of technology seen in a long while that can impact the end-to-end lifecycle of financial services. This is unlike previous technologies that had resulted in a narrower impact.
GenAI and its catalyzing of innovation will improve outcomes for product development, customer service, operational efficiency, and marketing:
1. Value realization with shorter lead time: GenAI can shorten the lead time for product development, leading to a faster time to market and a shorter time to realize value.
2. Faster, better decision-making: GenAI enables faster, broader access to large quantities of data, and sophisticated analytics, resulting in faster insight generation and decision-making.
3. Improved operational efficiency: Many functions will benefit from GenAI working alongside their teams to increase efficiencies through better knowledge search-and-retrieval, and various modes of content generation, amongst others.
4. Increasing accessibility of GenAI: GenAI is maturing quickly, and business leaders do not want to miss this moment to build capabilities and start moving GenAI into production environments.
That said, adopting GenAI is not without challenges. Business leaders will need to adopt the right strategic approach to AI infrastructure and keep their eyes on potential changes to legislations governing AI implementation and management.
DigiconAsia: How do business leaders overcome GenAI planning/adoption challenges, avoid potential pitfalls and tool up to evolve as legislation and emerging cyber consequences rear their heads?
DM: Firstly, identifying a suitable adoption pattern for Gen AI implementation is essential for firms to find the right offering for their unique needs.
Collaborating with experienced technology partners could be invaluable in navigating the complexities of this emerging technology.
Also, organizations will need a few quick wins — such as starting with available data and use cases that can lead to business impact — while also putting in place a platform and operating model for GenAI in production.
Finally, broad stakeholder involvement is also key. For example, compliance and risk officers will need to develop familiarity with regulatory frameworks and GenAI compliance issues so that appropriate policies (such as Responsible AI) can be implemented.
DigiconAsia: For the benefit of our readers, can you cite some real use cases for GenAI in the fintech and other industries?
DM: At the product conceptualization stage, GenAI is often used as a “co-pilot” for activities like understanding current user behavior, retrieving and summarizing existing and competitors’ product information, and creating suitable user personas and product requirements.
- During the development phase, GenAI is used across the Software Development Lifecycle from writing code, documentation, testing, and deployment.
- At the go-to-market stage, which is also integral in product development, GenAI is used for generating marketing content: text, images, animations for social media, press releases, etc.
- GenAI can reduce the cycle time for each of these activities and the friction between them, leading to real productivity gains.
- Customer service is also another aspect that has already started to reap the benefits of GenAI:
- ✓ When faced with complex customer enquiries, such as for complicated insurance products, call center operators can use GenAI to quickly retrieve detailed relevant information
- ✓ Professionally-worded and customer-oriented email responses to customer enquiries for many questions can be generated instantly even by not-so-experienced staff.
- ✓ Customer engagement cases can be summarized and analyzed quickly by GenAI for management to get a quick overview of what customers’ sentiments, feedback, and complaints are.
DigiconAsia: What do you think of the dire warnings about some AI dignitaries’ urgent calls for a global moratorium on overall AI development (including GenAI), given numerous uncanny/unpredictable behavior patterns in cutting-edge self-learning systems?
DM: Calls for the global moratorium tend to relate to Artificial General Intelligence (AGI), which is somewhat different in nature to the applications of GenAI we see in the financial services industry.
That said, there are certainly risks associated with using GenAI today in the industry. According to our own experts, one example of the risks of GenAI includes model hallucinations that can lead to customers being misled or confused. Addressing this issue requires:
✓ Rigorous testing and continuous monitoring, compliance with regulatory frameworks, and often humans-in-the-loop at the initial stages
✓ Testing and monitoring conducted as part of a broader platform & operating model (including GenAI governance and Large Language Model (LLM) Ops) that is crucial to get right.
✓ The above steps need to build on robust IT and data governance, to enable GenAI to be successful at scale and to reduce potential deployment risks. Specialized technology partners may need to be roped in for the necessary expertise to build and implement the necessary platforms, operating models, and end-user use cases, alongside that of the largest enterprises.
DigiconAsia thanks Dipen for sharing his fintech GenAI insights with readers.
