Not too long ago, it was impossible to withdraw money if a physical bank branch was closed. In the 1960s, this occurred to British inventor John Shepherd-Barron, who was left without access to his money after arriving just minutes after his London bank branch had closed. This got him thinking about a machine that could deliver money to customers at any time. Inspired by a chocolate vending machine, the ATM was born and the customer experience was changed forever.
Barclays Bank saw the opportunity presented by Shepherd-Barron’s innovation and installed the first ATM in 1967. Since then, we have seen technology enhance the customer experience in the banking sector in a number of ways. In fact, the sector has always been a pioneer in technological innovation. Continuing this trend, the financial industry today is looking closely at the opportunities offered by generative artificial intelligence (AI). However, its successful implementation requires both cutting-edge technology and good practices.
Generative AI use cases in the Banking sector
Today, there are many different types of AI, depending on the problem we want to solve. Generative AI is capable of generating content, text, images or sounds, from existing data (training data) and in response to the instructions we give it in the form of prompts. Some of the best-known examples are ChatGPT, Dall-E, Midjourney, GitHub copilot or Microsoft copilot.
Banking has been using AI for particular processes in its value chain for some time, such as analyzing risky operations, preventing fraud or predicting the behavior of securities on the stock market. These depend on specific AI tools, only useful for a particular need. However, generative AI has a competitive advantage in that it can be easily adapted to different organizational processes, from summarizing documents to proposing lines of code for software developers. This created a market for generative AI in the financial sector that was valued at $1.085 billion in 2023, with expectations it will reach $9.4 billion by 2032, growing at a CAGR of 28.1%.
According to a recent OECD report, generative AI in the banking sector is being used today to both improve productivity and for value creation. In the first case, its use is found in reporting processes, translations, human resources management, AML (anti-money laundering), fraud detection, and writing software code. However, its greatest potential lies in the possible generation of value for customers: the creation of new products, segmentation for targeted marketing, onboarding, authentication and customer service. Generative AI is expected to improve competitiveness in all banking verticals: consumer banking, financial services, payments, insurance, asset management, private banking and wealth management.
If ATMs revolutionized retail banking, generative AI has come to transform all financial businesses.
A new era in Customer Service powered by GenAI
Take the customer experience to the next level.
Major opportunities, minor risks
The implementation of customer-focused generative AI is experiencing uneven development in the banking sector, with fintech companies leading the way. Cleo’s slogan, “AI meets money,” is a transparent admission of its use of AI in managing its customer relationships. Similarly, Swedish company Klarna, which offers payment solutions for online shopping, demonstrated the remarkable performance of its customer-facing generative AI after just one month of operation, with Klarna’s AI assistant holding 2.3 million conversations – about two-thirds of all monthly interactions – carrying out the equivalent work of 700 agents across 35 languages and going on to resolve customer issues in two to 11 minutes, on average. The level of customer satisfaction was also similar to that achieved by a human agent. All in all, Klarna’s revenues are expected to increase by $40 million by 2024.
Other large financial corporations are in an exploratory phase with generative AI. Goldman Sachs is developing a dozen initiatives in generative AI, for instance, including writing lines of software code and generating documents. However, the company is being very cautious about applying generative AI to client-facing processes given the highly regulated nature of financial services. Morgan Stanley, meanwhile, is developing a chatbot based on ChatGPT to guide its wealth management clients regarding their investments. It is also expected that, with the customer’s permission, the chatbot will be able to create a summary of the conversation, compose a follow-up email suggesting the next steps, update the bank’s sales database, and schedule follow-up appointments. Insurance and financial services company John Hancock has implemented a Microsoft Azure generative AI solution in its user support center, managing to reduce service time and freeing up agents for issue resolution in more complex scenarios. In addition, part of the CAU team has been trained in low-code programming in the hope it will lead to an improvement in its services.
Admittedly, although generative AI’s journey in the realm of customer experience has undoubtedly begun, it is not without risk. For example, delivery firm DPD had to disable its customer support chatbot after it insulted a customer and criticized the company itself. The situation occurred because the affected customer deliberately forced such a situation, asking the chatbot to adopt a particular role. In any case, it revealed a security problem.
The AI Accelerator Institute has also highlighted other risks relating to the protection of personal data, customer privacy, intellectual property of training data, biased or erroneous responses, and the difficulty, sometimes impossibility, of determining the explanation of an outcome produced by generative AI. The latter is of particular interest to the financial sector, as it is strictly regulated and audited.
A well-governed Generative AI
To implement a customer experience-oriented generative AI solution capable of realizing its full potential while minimizing risks, an approach that integrates business, governance and technology is needed. With this vision in mind, we recommend the following:
- Start with use cases in very narrow environments and with well-defined initial functionalities.
- Develop a clear business case that includes opportunities and risks.
- Define guardrails that set ethical boundaries for generative AI.
- Establish OKRs that measure the success of any solution, including business, customer and ethical metrics.
- Determine to what extent the solution will be based on third-party or in-house developments.
- Have a well-prepared, quality and traceable training data set.
- Train everyone involved in the solution, from developers to customer service agents, on the risks of generative AI.
- Define the interaction channels with customers so they are informed that they are interacting with an AI and can use alternative communication channels if desired.
According to the IBM Institute for Business Value, 85% of executives believe that in two years, generative AI will interact directly with customers. Looking back at the origins of ATMs, to withdraw money, a bank employee had to provide the user with a special check, impregnated with carbon 14, before a unique PIN could be entered and money could be received. Today we interact with ATMs directly through our bank card or our cell phone.
No journey is straightforward. Innovation is simply a path. We still have to know how to walk it.
Take the next step in your GenAI journey
Eraneos is your perfect partner for integrating Generative AI into your service organization. Why? Because we’ve been around a long time, with years of experience solving Big Data problems with AI. Because our team is world-class, with over 150 experts who live and breathe artificial intelligence. And because we combine technology smarts with business know-how: you’ll find people who understand your challenges and want to solve them. People like you.
But above all, Eraneos helps you optimize that balance between performance and cost when it comes to Generative AI. Our service spans the entire AI journey: from initial strategy to project execution. We work with you to choose the right options and develop the integrations and applications that work for your situation.
Reach out to us today to explore the next steps in your GenAI journey!
