RBC researchers win global AI challenge for finance
RBC Borealis researchers placed first in the AI Agentic Retrieval Grand Challenge at the 2025 ACM International Conference on AI in Finance. The competition tested how well AI systems could locate and interpret relevant information buried in complex financial filings.
The work addresses a concrete problem in financial analysis: how to help AI systems find the right data in dense documents and deliver accurate, evidence-based answers. In finance, where precision matters and context shapes conclusions, this capability changes how analysis scales across institutions.
The technical problem
Financial filings contain thousands of pages of dense text. Analysts need to extract specific information quickly and accurately. The challenge was teaching AI to read like an expert analyst-identifying which details matter and which don't.
RBC's winning approach prioritized relevant content within large volumes of complex financial documents. The solution combined research with practical application grounded in real-world financial problems.
Why this matters for finance professionals
The win demonstrates that RBC has built research capability that works on actual business problems. Siqi Liu, Machine Learning Research Team Lead at RBC Borealis, said the team drew on challenges the bank was already exploring internally.
"That gave us a strong foundation going into the competition and helped us build on research that was already grounded in real-world financial problems," Liu said.
Eric He, Research Director at RBC Borealis, said the achievement reflects the bank's commitment to advancing AI for Finance in ways that strengthen operations and support future innovation.
The team
- Siqi Liu, Machine Learning Research Team Lead
- Amin Shabani, Machine Learning Researcher
- Mohammed Suhail, Senior Machine Learning Researcher
- Shuvendu Roy, Machine Learning Research Engineer
The result reflects how AI Research focused on specific business challenges can produce both academic recognition and practical tools. For finance teams, the implication is straightforward: better AI systems for document analysis mean faster, more accurate information extraction from the filings and reports that drive decision-making.
Your membership also unlocks:
