Financial Services
RBC Capital Markets Advances Research and Banking Capabilities With Agentic AI
Objective
RBC Capital Markets, in partnership with Borealis AI, has implemented advanced agentic AI technologies to enhance document processing, research capabilities, and banking services. By leveraging NVIDIA AI Enterprise, RBC has dramatically improved productivity and accelerated strategic goals across capital markets, focusing specifically on its Research and Banking divisions.
Customer
RBC Capital Markets
Use Case
Generative AI / LLMs
Key Takeaways
- Document processing capacity has increased 10x, with improved accuracy
- Research QuickTakes Report generation time is up to 60% faster
- Complex client queries are now answered in near real time, rather than taking hours or even days
- Alpha pattern discovery timeline is expected to decrease from 12 months to 2
Agentic AI for Real-Time Research and Increased Productivity
RBC Capital Markets set out to optimize how analysts deliver value to clients in a rapidly evolving and competitive environment. The team wanted to broaden coverage across sectors and produce faster, deeper insights for clients, but traditional manual processes made it difficult to accelerate productivity. Analysts often needed hours or days to parse information and build scenarios, leaving less time for more strategic and client-focused work.
To address these challenges, RBC is building agentic AI with NVIDIA AI Enterprise, enabling the firm to scale and enhance its Aiden AI platform across capital markets. RBC Capital Markets developed the AidenResearch Platform—a suite of specialized AI agents tailored for its Global Research division. This platform enables analysts to focus on high-value, data-driven client work while AI agents handle tasks such as generating earnings-related content, summarizing calls, surfacing insights from data catalogs, and providing real-time access to filings. As a result, the Research team has been able to deliver deeper insights to clients faster than ever before.
Building on this success, RBC is developing AidenBanker—a unified platform for bankers that connects essential functions like automated pitchbook creation, pre-meeting preparation, and comprehensive note-taking.
Other lines of business across RBC Capital Markets are now using the Aiden platform. With over 8,000 users benefiting from these capabilities, RBC is delivering timely, data-driven advice to clients, shortening time-to-market, and improving the overall client experience.
Using Advanced AI Frameworks and Technologies to Enhance Performance and Scalability
To support the growing application’s performance needs, RBC Capital Markets built an AI factory to adopt the NVIDIA Accelerated Computing platform. This AI factory is enabling use cases with hybrid infrastructure powered by on-premise and hybrid cloud solutions. Using an AI factory built with NVIDIA NeMo™ Agent Toolkit and NVIDIA NeMo™ Retriever, integrated with KX’s KDB.ai time-series-aware vector store, RBC bankers and financial analysts can provide real-time trading insights and market intelligence to its customers.
The NeMo Agent Toolkit enables automated agent workflows and large language model (LLM) queries, which is transformative for financial analysts and bankers using Aiden.
RBC deployed NVIDIA Nemotron™ models via NVIDIA NIM™ to ensure AI agents respond with fast, reliable reasoning and analysis, which is critical for financial analysts and bankers relying on Aiden for decision support. NIM offers streamlined deployment and scalable orchestration, enabling seamless agent updates with increasingly advanced models and maintaining low-latency analysis to uncover patterns and generate solutions in real time.
Measuring Aiden's Capabilities and Reach
Implementing NeMo Agent Toolkit and NeMo Retriever yielded significant performance gains and substantial cost savings while maintaining enterprise-grade scalability and compliance. Within just five months, NeMo Agent Toolkit drove tangible results and the highest degree of confidence, scaling Aiden to over 8,000 users across the bank.
Because this toolkit enables Aiden to handle complex client queries using agentic AI, answers are provided in near real time, significantly reducing the time burden on staff who previously spent hours or even days on these tasks.
Report generation time is now up to 60% faster, and this acceleration is expected to shrink the time for discovering alpha patterns from 12 months to just 2.
Using agentic AI built with NVIDIA AI Enterprise, and powered by an AI factory and NVIDIA Accelerated Computing, RBC is successfully scaling and enhancing its Aiden AI trading platform. The goal of AidenResearch is to enable analysts to cover up to 50 companies instead of the traditional 15, while AidenBanker aims to increase productivity among investment bankers.
Implementing these agentic AI platforms supports RBC's ambition to achieve C$700 million to C$1 billion in enterprise value generated from AI-driven benefits. This strategic expansion significantly reduces manual efforts, creates better user experiences, and ultimately improves client services through faster, more accurate insights and decision-making capabilities.
“AI is one of the most transformative technologies impacting the world today, and we’re scaling our innovation engine to transform the way we work, think, and deliver for clients. Our work with NVIDIA across AidenResearch and AidenBanker has been powerful in generating customized generative AI solutions to unlock talent productivity and deepen engagement through more data-driven client experiences.”
Bobby Grubert
Head of AI and Digital Innovation, RBC Capital Markets
Discover how NVIDIA NIM can accelerate AI workflows in financial services.