In Hong Kong, the adoption of generative artificial intelligence (genAI) in the financial sector remains focused on internal applications aimed at automating tasks, improving workflows and sharing knowledge.
According to research by the Hong Kong Monetary Authority (HKMA), KPMG China and Quinlan and Associates, 80% of genAI use cases in the financial services sector are centered on these internal functions.
The study, which polled 137 industry practitioners and conducted in-depth interviews with 16 organizations across the technology, banking, securities, and insurance sectors, found that concerns about the technological and regulatory maturity of genAI are limiting its broader adoption, particularly in customer-facing roles.
Most of the financial institutions interviewed expressed reservations about whether genAI is sufficiently advanced to consistently deliver reliable and accurate results directly to customers, with one technology provider sharing that their financial institution clients expect improved technological maturity before adopting genAI solutions.
Industry stakeholders have warned of the risks of genAI. The International Monetary Fund (IMF) highlighted in a 2023 report that while the technology offers potential benefits like improved efficiency and enhanced customer engagement, it also poses challenges. These include risks related to embedded bias, privacy issues, cybersecurity threats, and threats to financial stability. Until regulations evolve, human supervision and enhanced oversight are necessary to mitigate these risks, the IMF said.
How Hong Kong financial institutions are using genAI
The study and interviews identified a total 59 genAI use cases, covering both current and planned adoption scenarios in Hong Kong’s financial services industry. Of these, 47 (or 80%) are internal applications aimed at improving operational efficiency.
The main purposes of these applications are to address three major pain points:
- Information overload: GenAI is supporting Hong Kong financial institutions in processing large volumes of data and condensing it into clear and concise summaries;
- Time-consuming, repetitive tasks: GenAI is used to streamline resource-intensive workflows and automate repetitive and lower-value tasks such as idea generation, code generation, outreach message customization, and report drafting; and
- Human errors: GenAI is used to deliver consistent, high-quality outputs, and reduce the likelihood of errors.
Among the 59 use cases, eight were identified as general applications involving tasks such as automating meeting minutes and generating code, while the remaining 51 use cases were specifically relevant to the financial services industry.
Among these, use cases across the service fulfillment value chain were found to be the most prevalent, with research and analysis, in particular, emerging as the top area for deployment. This is evidenced by a total number of ten financial institutions reporting using the technology to automate internal and external research efforts, including synthesizing data from multiple sources, automating report creation, and providing real-time updates to product reports – the largest number across the 50+ use cases.
Another key area of genAI application in the financial services industry is customer servicing, where seven financial institutions indicated using the technology for enquiries management and complaints analysis, making it the second most prevalent use case. In these scenarios, genAI is used to generate personalized responses quickly based on customer data, enhancing both engagement efficiency and customer retention. The technology also enhances complaint analysis and handling capabilities, with one institution reporting a 50% reduction in complaint analysis time.
Finally, genAI is also being leveraged for operations and risk management, specifically in risk profiling and customer due diligence. Four institutions indicated using the technology to enhance risk assessments, allowing employees to profile client risks more quickly and accurately by summarizing customer information and identifying relevant details for further investigation. By automating the document review process, three interviewees highlighted their ability to arrive at faster and more accurate due diligence decisions, with a global credit rating agency citing a 27% reduction in time needed for credit analysis tasks.
Boosting genAI adoption
The HKMA’s research paper on genAI is part of a broader effort to promote fintech adoption in the city, with AI being a key area of focus. In August 2024, HKMA, in collaboration with the Hong Kong Cyberport Management Company Limited (Cyberport), launched the GenAI Sandbox, allowing financial institutions to develop, test and pilot genAI technologies in a risk-controlled environment.
Initially, these trials will focus on enhancing risk management, anti-fraud measures, and customer experience. Participants will receive early supervisory feedback and guidance from HKMA, the central bank said.
For HKMA, the sandbox will allow it to monitor the development and implementation of these AI technologies, ensuring robust risk management practices before full market deployment.
GenAI is expected to have a significant impact across all industry sectors but banking will be among those the most impacted, according to McKinsey and Company. Estimates by the consultancy show that the technology could potentially generate value from increased productivity of 2.8-4.7% of the industry’s annual revenues, translating to an additional US$200-340 billion in revenues annually.
Featured image credit: edited from freepik