Beyond the buzzword: applications of Generative AI across Banking and Financial Services

Throughout history, radical new ideas have always been the catalyst for profound change. From a technology perspective, the events that have shaped the world the most are the invention of an innovative new solution.

In the early 1980s, the launch of the internet and the ensuing era of big data changed the state of play for global business forever. Today, it seems we are living through another pivotal moment, with the rapid acceleration of development of artificial intelligence driving a seismic, cross-industry shift.  

Looking back, the onset of digitalized services brought on the era of big data, enabling companies to gather information about consumers’ preferences, implement analytical capabilities, and transform their business approach using rich, data-driven insights. Since then, further advancements in technology, such as diverse forms of AI, rule-based systems, low-code tools, and machine learning algorithms, have been continuously added to tech stacks, sharpening these organisations’ ability to maintain a competitive edge.

On top of this, companies have also enhanced access to data and harnessed the power of Business Process Optimization tools (BPOs), revolutionizing the way of working, creating operational efficiencies and boosting the overall employee and customer experience.

Despite all of these advancements, we still lack the ability to communicate with machines in a “human way”. The emergence of generatative AI – the next sea change for the business world – might just be the tool to take us there. 

Generative AI: more than a buzzword

We might see the first significant breakthrough in deep learning to be the publishing of  Attention is All You Need by Cornell University in 2017. However, it wasn’t until late 2021 when Generative AI, a type of automated learning technology capable of creating new written, visual, and audio content from existing data, changed the game.  

It has demonstrated astonishing capabilities of analyzing and producing new applicable and scalable information to different business use cases using deep learning techniques, neural networks, and other advanced algorithms. Companies in various sectors, including finance, healthcare, and retail, have incorporated Generative AI into their workflows, creating cutting-edge solutions and disruptive value propositions.  

Generative AI has taken the assistive technology of machines to a new level by allowing us to “chat with our data,” unlocking powerful capabilities for both technical and non-technical users and reducing the development time of applications across organizations. In the spirit of going beyond the buzzword, we have set out to identify and describe the different use cases in which processes and operations from the banking industry can benefit.

Uses cases in Banking

As expected, there are numerous use cases that apply directly to the banking Industry.  At a sales and customer service level, Robot Process Automation (RPA) can embrace new capabilities through a more customized approach to natural language. On the operational side, decision-making automation and a newly upgraded set of BPOs can streamline more complex tasks, increasing productivity and competitiveness. Generative AI even has a role to play in filling talent gaps, assisting with legacy code modernization and development to compensate for scarcity of highly technical roles in IT.

Unlocking sythentic data

A customer’s data, such as their financial information, is incredibly sensitive and thus requires banks and financial institutions to safeguard it at all costs to avoid data breaches and adhere to GDPR and other regulations.

At the same time, this data is also vital to building the AI models that underpin processes like virtual financial advisors or portfolio optimization tools. These algorithms “feed” from large amounts of information to identify statistical patterns, learn, and replicate critical attributes.

With the production of synthetic data, data that closely mimics accurate information without using sensitive information such as the name or account number of a banking customer, corporations can provide the quantity and variety of data necessary to train and finetune these models, allowing predictive analysis and the definition of BI, AI rules.

For example, the robust neural networks of Generative AI can replicate and tabulate information like a historical list of financial transactions, enabling behavioral analysis to design customized products and services for each banking customer.

Blockers for Generative AI

As a statistical tool trained from vast amounts of data, an AI model is highly dependable on the quality of information provided; if given the wrong input, it could process non-factual details, learn from it, hallucinate, and output inaccurate answers. To avoid this, AI must incorporate principles, regulations, and standards that encompass an ethical AI strategy.

Whilst of course it has the power to be transformational, Generative AI is still not a human brain; it does not have reasoning capabilities, which means decision-making must always be supervised.  Thefeore, the responsibilities still have to be held by humans. Technology can be a useful copilot, but humans must always remain in the loop.

Our vision for the future

The Generative AI journey has become more accessible due to the availability of pre-trained models; nevertheless, as technology advances rapidly, organizations must define a strategic approach to integrate AI into their daily operations effectively.

Embedded in the Digital Technology unit, we have recently opened an all-new AI HUB in Valencia. Its main objective is to develop AI by design, which means:

• Establishing the right ecosystem of partners, users, and platforms
• Strengthening AI adoption with a clear value proposition
• Adapting the technology to the culture and values of the organizations
• Transforming traditional productive models to implement new initiatives.

To summarise our vision in a sentence, we believe that the real challenge with AI is not about the technology itself, it is about articulating the right deployment strategy, organization, and governance.