Overview:
RBC Capital Markets (RBC), in collaboration with Borealis AI, deployed an agentic-AI platform (branded “Aiden”) built on NVIDIA technologies (including NVIDIA AI Enterprise, NeMo, NIM, RAPIDS) to accelerate document processing, research content creation, banking pitch-book generation and client engagement across its Global Research and Banking divisions.
The problem targeted manual, time-consuming workflows: analysts spending hours or days parsing filings, writing content and preparing materials, limiting coverage depth and slowing response to client queries.
The solution operates in the finance domain (research, investment banking) and uses AI to process large volumes of structured and unstructured data (earnings calls, filings, data catalogs, market time-series). The platform scales to over 8,000 users and supports discovery of alpha patterns, real-time query responses and streamlined content production.
Key Features:
Ingests and processes large volumes of documents (earnings calls, filings, pitchbooks) to accelerate content generation.
Summarises and generates research “QuickTakes” reports, reducing time to insight.
Answers complex client queries in near-real-time via AI agents, replacing hours-or-days manual work.
Discovers alpha patterns by analysing time-series and vectorised data stores (e.g., KDB.ai) tied to financial markets.
Automates banker functions via AidenBanker: pitch-book creation, pre-meeting prep, note-taking and integration of research.
Employs a hybrid infrastructure “AI factory” for on-premise and cloud GPU-accelerated computing to scale enterprise deployment.
Results & Impact:
Document-processing capacity increased by 10× while improving accuracy.
Report generation time reduced by up to 60 %.
Alpha-pattern discovery timeline shortened from 12 months to 2 months.
Over 8,000 users now using the Aiden platform across RBC’s business units.
Strategic value generated (target) between C$700 million and C$1 billion from AI-driven benefit.
Analysts broaden coverage: goal to cover ~50 companies instead of ~15.
AI Technology:
AI Model Types: Large Language Models (LLMs) + Agentic AI + Vector retrieval models + Time-series aware vector stores.
AI Purpose: Automate document ingestion, summarization, query-response, insight discovery.
Application Type: Finance – Research & Investment Banking (Operations, Client Service, Insight generation).
Target Users:
Research analysts
Investment bankers
Client-engagement teams
Market-intelligence teams
Data science/AI operations staff
Sources:
https://www.nvidia.com/en-us/customer-stories/rbc-capital-markets