Hong Kong Monetary Authority

04/02/2026 | Press release | Distributed by Public on 04/02/2026 02:24

Opening Remarks at FiNETech7 on “Data Excellence for Intelligent Risk Management”

Opening Remarks at FiNETech7 on "Data Excellence for Intelligent Risk Management"

Speeches

02 Apr 2026

Opening Remarks at FiNETech7 on "Data Excellence for Intelligent Risk Management"

Carmen Chu, Executive Director (Banking Supervision), Hong Kong Monetary Authority

  1. Good morning, everyone. It's a pleasure to welcome you all to FiNETech7, the seventh edition of our flagship series on Fintech promotion.
  2. Today, we will focus on data excellence. You may wonder why are we putting a renewed emphasis on "data"? Isn't it something the industry has been working on for years?
  3. The answer is simple. Data now presents a new set of opportunities for intelligent risk management.
  4. Over the years, financial institutions have established strong data governance, especially for core, structured data. Meanwhile, the adoption of generative artificial intelligence (GenA.I.) and other new technologies require a tremendous amount of unstructured data to power them.
  5. To keep up with the demand, we have to move past the stage where data is scattered across multiple internal systems. And it's no longer appropriate or adequate to rely on manual workarounds.
  6. A recent study1 shows that 86% of business leaders identify data as their most underutilised asset. The opportunity now is to truly utilise data. By doing so, we can enhance consistency, raise effectiveness, and facilitate sound decision-making. Now is the time for us to actively explore the full potential of data through our new Risk Data Strategy.
  7. Under the new strategy, data excellence can't be just a compliance concept. It must be a competitive advantage. To guide our journey, let me introduce a framework called DataGRID. It stands for: granularity, reliability, ingenuity, and discoverability.
  8. First, granularity. We need data at the right level of detail - whether at the transaction, collateral, or position level. It allows us to slice and dice information, giving us the context needed to identify risks without losing time to manual processing.
  9. Second, reliability. Data is useless if it's not accurate. Reliability is about building a consistent, auditable single source of truth. After all, the advanced artificial intelligence (A.I.) models that we want to build are only as good as the data feeding them. Reliability ensures that results and recommendations are built on facts, not flaws.
  10. Third, ingenuity. This is about designing our data for the future. It means building the capability to extract insights from both structured and unstructured data. Ingenuity empowers us to leverage A.I. to detect hidden anomalies, predict market shifts, and generate forward-looking risk strategies.
  11. And lastly, discoverability. Think of this as a search engine. While market dynamics may shift overnight, we should not have to spend weeks hunting for the right data. Discoverability allows unified, instant access to critical information - spend less time searching for data, and more time acting on it.
  12. This DataGRID framework defines the most important attributes of data excellence: from governance to technology and culture. The ambition is to transform data into true intelligence.
  13. The next question is "how". We recognise that transformation can't happen in a vacuum. We must therefore work together as guided by the DataGRID framework, with three core building blocks.
  14. The first building block is co-design. This collaborative approach is especially critical for driving reliability and ingenuity. By working hand-in-hand across the ecosystem, we can be sure that our shared data foundation is built on real-world experience - accurate, robust, and fit-for-purpose.
  15. The second building block is a reference architecture. This serves as a powerful enabler for discoverability. Together we will co-design a reference architecture tailored to the unique data environment of the financial sector, to allow for the seamless integration of structured and unstructured data, both from internal and external sources. This will help individual institutions to enhance data pipelines and build forward-looking risk models.
  16. The third building block is Granular Data Reporting, or GDR. As the name suggests, this naturally focuses on granularity. Some of you may have been involved in the first phase, which piloted the collection of granular, exposure-level data from 50 banks. Then GDR 2.0 has successfully replaced the template-based residential mortgage lending survey since last year.
  17. Today, building on that success, we're launching GDR 3.0 - a multi-phase, multi-year revamp guided by the principle of "Report Once, Use Multiple". We're streamlining or discontinuing obsolete surveys and returns, ingesting more relevant data, and thus enabling more intelligent risk management.
  18. In fact, our supervisory team has already trained our first machine-learning model using granular corporate loan data, exploring the possibility of predicting potential credit migration.
  19. Looking ahead, the development of GDR 3.0 will also pave the way for on-demand data reporting. Risk does not operate on a neat, month-end schedule, and our data collection shouldn't either. As a longer-term feature, we'll explore the possibility of doing away with rigid reporting cycles. By leveraging straight-through submission, data can be transmitted seamlessly in response to sudden, ad-hoc market events.
  20. Through these three building blocks, we are collectively establishing an industry-wide DataGRID - underpinned by strong governance and a shared commitment to data excellence. And with this foundation, we'll be advancing towards more intelligent risk management and supervision.
  21. Before we start, I would like to extend my sincere gratitude to our co-host, Cyberport, and to our supporting organisations: the Securities and Futures Commission, the Insurance Authority, the Mandatory Provident Fund Schemes Authority, the Authority of Qianhai Shenzhen-Hong Kong Modern Service Industry Cooperation Zone of Shenzhen, The Hong Kong Association of Banks, the Hong Kong Science & Technology Parks Corporation, and the FinTech Association of Hong Kong, as well as to KPMG and Quinlan & Associates for facilitating the arrangements.
  22. I would also like to thank our distinguished speakers for sharing valuable insights with us today.
  23. The journey to data excellence begins not with technology but mindset. Supported by DataGRID, let's co-design GDR 3.0 to excel in data-driven, intelligent risk management and beyond.
  24. And the journey to data excellence begins now. Thank you.

1 "Intelligence at scale: Data monetization in the age of gen AI" published by McKinsey on 31 July, 2025.

Hong Kong Monetary Authority published this content on April 02, 2026, and is solely responsible for the information contained herein. Distributed via Public Technologies (PUBT), unedited and unaltered, on April 02, 2026 at 08:24 UTC. If you believe the information included in the content is inaccurate or outdated and requires editing or removal, please contact us at [email protected]