How generative AI is redefining finance

GENERATIVE artificial intelligence (generative AI) is progressing so fast that the ChatGPTs and Google Bards of the world feel like yesterday’s news. Beyond text-based use cases, we now have generative AI tools that produce codes, audio, video and 3D models with a few direct, well-articulated prompts. There’s no doubt that AI is quickly becoming an integral part of our workflow.

The technology represents a game-changing breakthrough in the value that enterprises can create.

According to McKinsey, generative AI could add over US$4 trillion to the global economy and US$8 trillion in productivity gains. With a myriad of models and tools readily available, some even for free, enterprises must already be generating tremendous business value even as we speak – in theory, at least.

Before generative AI can reach its fullest potential, businesses first need to get the data sprawl under control. For the banking, financial services and insurance (BFSI) industry notably, data management is at the core of a deeper, more valuable generative AI integration.

Ripe for an AI revolution

Generative AI works by learning patterns and structures from a large data set and then using this knowledge to create new, previously unseen data. Even though our journey with it is only the beginning, there are already high-value use cases within the BFSI industry, specifically in risk management, customer engagement and fraud detection.

In risk management, generative AI can identify trends and predict market movements by analysing historical market data, news articles, social media sentiment and economic indicators. This can help banks and investment firms make better, more informed investment decisions.

Generative AI can also incorporate a broader range of information, including non-traditional credit data such as social media activity and transaction history, to create more accurate credit risk assessments. More holistic credit profiles allow lenders to make fairer lending decisions and even improve financial inclusion among unbanked populations.

On the customer engagement front, many BFSI firms are already using generative AI-powered chatbots and virtual assistants to handle routine inquiries and even provide financial advice. Generative AI algorithms can also be used to analyse customer transaction history and behaviour, and then provide personalised product recommendations. By deepening the understanding of customer needs and expectations, businesses can respond with more authentic and relevant interactions to vastly improve customer engagement and loyalty.

Finally, considering the rise of cybercrimes, generative AI can identify unusual patterns and behaviours in transaction data, thus helping BFSI companies detect and prevent fraudulent activities in real time. This is crucial for ensuring the security of customer accounts.

Reining in the data sprawl

While the potential of generative AI is immense, the AI itself is only as good as the data fed to it. The challenge is that, over time, many enterprises have grown their tech stacks and accumulated massive tech debts and unmanaged data sprawl. As such, enterprises sometimes end up spending millions just to move and replicate data across numerous pipelines and silos on legacy systems. This is an inconducive, even hostile, environment for effective generative AI.

This is where data management comes in. Generative AI models work only if there is high-quality, well-structured data available. Proper data management, then, helps to ensure that the data that BFSI companies have amassed is clean, reliable and free of errors.

This is especially critical considering that BFSI firms are expected to be the authorities on market predictions and recommendations. Good data management, then, is not only a good practice, but also a matter of reputation and positive business outcomes.

Furthermore, one of the greatest obstacles to effective AI deployment is siloed data sources – and BFSI firms are rife with data scattered across different departments and systems. In this case, effective data management helps to integrate disparate sources and create a standardised, unified and comprehensive data set. This provides a holistic view of customer behaviour, transaction history and other relevant information, making it more valuable for generative AI models to train on and analyse.

Finally, given the sensitivity of financial data handled by the BFSI industry, encryption of data both in storage and in transit is crucial. This ensures that even if unauthorised access occurs, the data remains unreadable. Continuous monitoring and auditing of data access and activities must also be in place to detect and respond to security breaches or suspicious activities.

These practices not only facilitate the training and use of generative AI models, but also safeguard sensitive customer data, enhance regulatory compliance and support accurate decision-making within the industry.

In 2022, humans created, captured, copied and consumed about 97 zettabytes worth of data. If done right, each data point has the potential to generate massive amounts of value for businesses across all industries.

Navigating the complex and ever-changing landscape of financial technology is challenging enough for the BFSI industry. Embracing these technologies and practices is not merely an option; it is the path forward in ensuring the industry remains relevant, trustworthy and at the forefront of innovation in the years to come.

The writer is executive vice-president, international, at Teradata