AI Data Manager (FAF.ai) for Index Calculation

The goal was to automate FAF calculations for the Hang Seng Family of Indexes using generative AI, reducing manual effort, accelerating financial data processing, and improving the analysis of complex corporate structures while maintaining high accuracy.

2025

This use case aiming to enhance productivity in the index calculation process and expand data analytics capacity. This capability is integrated into the index calculations of the flagship indexes and more.

Impact Highlights:

• Enhanced Operational Efficiency: Lead time for FAF calculation has been reduced by two-thirds, allowing for more frequent client updates.

• Expanded Data Processing Capacity: The solution enables a 4X increase in data processing capacity, positioning the company to handle a more complex and dynamic capital market.

• Reduced Human Workload: Automation has significantly decreased manual effort, allowing the team to focus on product refinement.

• High Accuracy Maintenance Journey: Accuracy now stands at 90%+ with ongoing optimisation.

Cloud and On-premise environment

LLM, Python, PyTorch, TensorFlow, ONNX, CUDA

The project does not use solutions from the Global CIO catalog

Hang Seng Indexes Company partnered with INF, a Cyberport community member, to co-develop Gen-AI solutions for calculating the free-float adjusted factor (FAF)—a critical metric representing the percentage of publicly tradable shares of a listed company. INF integrated neuro-symbolic technology into its AI large language model to ensure the precision and logical consistency required for complex financial data.

Hang Seng Indexes Company partnered with Cyberport Community member INF Hong Kong to implement generative AI in the calculation into the Hang Seng Family of Indexes, including the Hang Seng Index, Hang Seng China Enterprises Index and Hang Seng TECH Index.

The “AI Data Manager (FAF.ai) for Index Calculation” project began in 2024 as a collaboration between Hang Seng Bank and Cyberport, launching the “Index Innovation Lab” to drive technology adoption in index operations. In November 2025, Hang Seng Indexes Company and INF Hong Kong (INF) have successfully transitioned Gen-AI from a laboratory concept to the adoption of Gen-AI generated data for the calculation of Hang Seng Indexes Company’s flagship indexes—the core indexes of Hong Kong’s capital markets. Today, this capability is integrated into the index calculations of the Hang Seng Family of Indexes—including the flagship Hang Seng Index, Hang Seng China Enterprise Index, and Hang Seng TECH Index.

This initiative has strengthened operational efficiency and increased data-processing capacity, while maintaining a high level of accuracy.

The successful deployment demonstrates Hang Seng Indexes Company’s leadership in technological innovation and reinforces its commitment to excellence in index operations.

Internal usage for solution itself, and the output of solution is generally applied to the calculation of Hang Seng Family of Indexes.