AI in Financial Services: From Internal Efficiency to Hyper-Personalization – What Institutions Are Doing

At Nordic Fintech Week 2024, the role of artificial intelligence (AI) in reshaping the financial services industry was a central theme. In my previous article, we explored the importance of board oversight, governance, and risk management in navigating this transformation. Now, it’s time to dive deeper into what financial institutions and other actors in the industry are actually doing with AI.

This article will focus on some of the real-world applications that financial institutions were sharing at the conference, from the internal use cases designed to streamline operations, to the hyper-personalization of customer experiences powered by AI. How companies are leveraging AI to improve decision-making, drive efficiency, and deliver customized services at scale.?

From AI-powered fraud detection to automated financial advice, we’ll take a look at what was presented and already in production, and the key learnings from their journey so far. As these institutions continue to evolve, their AI initiatives provide valuable insights into the future of financial services, and how technology will continue to shape the industry.?

As mentioned in the previous article, the thoughts put together in this article are based on my notes and interpretation and filtered through what I had time to experience and my personal biases and opinions.

1. How Financial Institutions Are Using AI

The presentations and interviews from Nordic Fintech Week 2024 reveal a range of AI applications across different institutions - I had the pleasure to be in the room when many presented, but by far not all:

• 花旗 and Ronit Ghose emphasizes the transformative potential of AI in finance, outlining various use cases such as coding and software development, customer service chatbots, investment research, transaction monitoring and compliance, credit risk and underwriting, and asset and portfolio management. They predict the impact of AI on these tasks will increase over time, leading to new products and business models. However, they also point to historical examples like the introduction of ATMs and spreadsheets, arguing that technological advancements in finance haven't necessarily resulted in immediate or significant workforce reductions.
• SAP and David Pontoppidan highlights the importance of AI maturity and the need to move beyond isolated projects to enterprise-wide AI strategies. They acknowledge the challenges of GenAI, particularly its probabilistic nature and the risk of project abandonment due to the hype surrounding it. They stress the need to focus on contextual, critical, and clear AI that integrates seamlessly with existing business data and processes. SAP presents its Business AI platform as a solution, emphasizing its contextual understanding, reliability, responsibility, and focus on intelligent automation.
• íslandsbanki and árni Geir Valgeirsson showcases a comprehensive AI roadmap that spans customer engagement, operational efficiency, and language preservation. They are exploring AI-powered chatbots with voice control, sentiment analysis for customer service requests, automated testing and documentation, fraud analysis, and even a financial advisor. Interestingly, they are collaborating with Microsoft and the Icelandic government to use AI, specifically GPT, to preserve the Icelandic language. They also acknowledge the importance of addressing risks associated with AI, including reputational damage, privacy concerns, data quality issues, operational risks, security vulnerabilities, bias, and the explainability of black-box models. Their Generative AI framework aims to mitigate these risks through robust data governance, prompt engineering, and model monitoring.
• Danske Bank and Christian Bornfeld positions generative AI as a key driver of its Forward ‘28 strategy, aiming to create a more personal, convenient, holistic, and proactive banking experience. They have invested significantly in AI, including an internal GPT solution called Danske GPT. Their use cases span various areas, including auto-reply, customer support chatbots, and task automation. They acknowledge the operational risks and ethical dilemmas associated with AI, emphasizing the need for experience and dialogue to address them. Danske Bank also highlights the evolution of its FinTech partnership model to accelerate the execution of its strategy.
• Danica Pension and Jesper B. focuses on the transformative potential of AI in digital healthcare. They believe AI can revolutionize healthcare by predicting and preventing illness, increasing efficiency, and enabling personalized treatment plans. They highlight Denmark's high adoption rate of digital health technologies and the EU's supportive regulatory framework for data sharing. Danica Pension's presentation demonstrates how AI can be leveraged to improve customer outcomes in a data-driven, personalized manner, a concept that can be applied to financial services as well.
• Topdanmark and Kasper Tj?rntved Davidsen presents TopGPT, a mixed-model AI chatbot that combines general knowledge with insurance-specific expertise. This approach enables TopGPT to handle a wider range of customer inquiries and provide more comprehensive answers. Topdanmark emphasizes its commitment to a Digital First strategy, aiming to answer 80% of customer inquiries within 20 seconds. They highlight the significant improvement in customer satisfaction, achieving a 10x higher NPS score with TopGPT compared to its rule-based predecessor. Topdanmark is continuously innovating with AI, developing TopGPT 2.0, which allows customers to access personalized information based on their insurance policies. They see GenAI as a key enabler of new customer experiences, business models, and revenue-generating opportunities.
• Lunar and K?re Kjelstr?m is also heavily investing in AI, particularly in hyper-personalization. They use AI to enrich customer transactions with additional information, create personas based on user data, and provide personalized financial advice. They are gradually automating various aspects of the customer experience, including onboarding, customer support, and task automation, while prioritizing customer safety and regulatory compliance. Lunar's CTO envisions a future where AI becomes a highly personalized financial advisor, capable of managing finances and making investment decisions on behalf of the customer, but acknowledges the need for a phased approach to address potential risks and ethical concerns.

2. Danish FSA and Key Messages on AI

The Danish FSA takes a cautious yet optimistic stance on AI in finance.

• They acknowledge the "huge potential" of AI but stress that it "does not replace good old-fashioned governance and risk management". They advocate for simpler regulations that are easier to understand and apply, particularly for smaller players. Their report on AI usage in the financial sector reveals that AI is "almost exclusively" used internally as a "supportive tool," with limited customer-facing applications. They attribute this limited adoption to several factors, including a lack of skills and resources, regulatory uncertainty, data quality concerns, and anxieties about customer acceptance.
• Across all sources, there is a clear consensus that AI is rapidly transforming the financial sector. The focus is shifting towards enhancing customer experiences through personalized services, faster responses, and 24/7 availability. However, there is also a strong emphasis on the need for a balanced approach, ensuring responsible AI development and deployment while addressing potential risks and ethical implications.

3. Key Learnings

Several key learnings (this is definitely not the entire list) emerge from the presentations and interviews:

• Early adoption is crucial: Companies that are slow to adopt AI risk falling behind. Early adopters are already realizing significant benefits, such as increased efficiency, improved customer satisfaction, and new business opportunities.
• A mixed-model approach is becoming the standard: Combining general AI with domain-specific expertise is key to creating more powerful and effective AI solutions.
• Risks must be addressed proactively: AI is not a silver bullet. Institutions need to invest in understanding and mitigating potential risks, such as data bias, hallucinations, security vulnerabilities, and ethical considerations.
• Regulation needs to adapt: Current regulations may not be agile enough to keep pace with rapid AI advancements. Simpler, more principles-based regulations are needed to foster responsible innovation and prevent overregulation.
• Cultural change is essential: For AI to be truly successful, organizations need to foster a data-driven culture and ensure that employees understand and embrace AI.