Generative AI in banking: From buzz to business value

Empowering innovation in today’s digital landscape

Swiss private bank Julius Baer has never shied away from new technology – in fact, they have been putting in significant investments into AI innovation. 

Launchpad, the bank’s innovation lab, was set up in Singapore to develop and test disruptive solutions. Launchpad’s focus is on experiment-driven approaches to develop and test disruptive solutions faster – with promising results in key areas like digital assets, decentralised finance, and generative AI. 

In the rigorously regulated landscape of the financial sector, a strategic, co-creation approach ensures that banks like Julius Baer remain on the cutting edge, responsibly harnessing the benefits of technological advancements while expertly mitigating risks. 

An agile outsider approach

Launchpad set out to achieve an ambitious long-term vision: to augment bank staff with AI-powered capabilities and to experiment with LLMs as a new disruptive technology stack for Julius Baer. 

As the bank’s strategic innovation catalyst, Launchpad plays a key role in swiftly validating proofs of concept and developing prototypes. This rapid innovation approach enables Julius Baer to effectively validate before scaling solutions across the entire bank, leveraging Launchpad's ability to transform innovative ideas into reality. 

“Julius Baer and Zühlke share a lean mindset that enables us to identify what doesn't work quickly and pivot effectively. Pivoting isn't just about making changes; it's about making informed decisions based on feedback and learnings, ensuring that we deliver real value as a team,” explains Anja Baer, Partner at Zühlke.

“Whether it’s experimenting with a novel technology like generative AI or bringing in creative minds to elevate user experience, data science specialists or tech expertise, we can swiftly mobilise the right capability or talent, allowing us to pivot and learn along the way,” Anja explains.  

“Working collectively has enabled faster implementation to internalise capabilities and resources together at one table,” Jonathan highlights. 

Generative AI in banking: Balancing innovation and compliance

As a ground-breaking technological advancement, generative AI has not only captured global attention but also emerged as a focal point for innovation in the banking sector. Recognising this, Julius Baer took a strategic step to explore its potential use cases in the banking space. 

The power of generative AI lies in its ability to address some of the most pressing issues in banking today. From the automation of repetitive tasks to enhancing knowledge search, generative AI presents a suite of solutions that can revolutionise how banks operate.  

On the flip side, the adoption of generative AI comes with its own unique challenges, particularly in the realms of cybersecurity and regulatory compliance. These are critical considerations that leading banks like Julius Baer must navigate to ensure both innovation and responsibility. 

Launchpad joined forces with Zühlke to tackle these concerns proactively by co-creating a proof of concept. This initiative demonstrates Julius Baer's commitment to innovation and excellence, showcasing how generative AI can seamlessly and safely integrate into the bank's operations. The proof of concept needed to: 

• Illustrate tangible impacts on selected use cases and
• Lay down a strategic framework for responsibly adopting this technology in the future

Launchpad’s innovation team partnered with Zühlke’s engineering and data science capabilities to co-develop use cases and propositions rapidly in an iterative manner. This collaborative, agile approach meant that the team was able to build and test lean prototypes quickly to solve the data domain challenge when it comes to generative AI. 

The ultimate goal was to support Julius Baer in aligning cutting-edge technological advancement with key business use cases. 

Try, fail pivot and learn

With only eight weeks of iteration, Zühlke and Launchpad debuted the final proof of concept: an AI-powered solution that deciphered publicly available knowledge bases established on plain-language questions. 

Early in the development process, the team determined that the solution would be most valuable for two specific use cases, with the aim of reducing downtime through focused research.

1. First, the solution helps find specific passages in thousands of pages of regulations - drastically reducing research downtime for compliance teams.
2. Second, the ability to use plain language to search through vast data lakes, minimising the time it takes to identify and summarise key data points for content creation.

The solution laid the infrastructure foundations for generative AI innovations in the future – ensuring that as the bank's AI capabilities grow, they remain secure, compliant, and aligned with the bank’s ethos of trust and innovation. 

“Our innovation approach is to validate quickly our hypothesis with a pilot group of users to generate real world feedback that will inform the organisational impact of such innovative ideas,” explains Jonathan.

“Rapid exploration and learning is the key to driving innovation but doing so in a safe environment where our security requirements are in place is something that we are very conscious of as well.”  

Co-creating the future of banking powered by AI

For financial institutions like Julius Baer, trusted advisers like Zühlke help accelerate and de-risk their digital initiatives, by offering complementary expertise and services that consistently add strategic value to their innovation efforts – from generative AI to future technologies over the horizon. 

Zühlke's global data and AI practice brings over a decade of deep domain experience in highly regulated sectors such as financial services, covering specific industry use cases in various business sectors. “Our clients possess deep expertise in their respective industries, and we mirror their commitment to excellence in the realm of technology,” Anja says. 

In this disruptive business environment, the bottom line of innovation is an ongoing process of learning and adapting. This mindset of ecosystem innovation with partners – which embraces uncertainty and constant change, is fundamental in ensuring technology investment delivers commercial and customer value.