RBC Capital Markets: From AI experimentation to front-office transformation

“AI at RBC is real. It’s making an impact.” — Bobby Grubert, Head of AI and Digital Innovation, RBC Capital Markets

Recently named North America’s best bank for research by Euromoney, RBC Capital Markets supports 22,500 clients across 100 countries with innovative products and services, forward-looking advisory, and excellence in execution. It’s renowned for global expertise and data-driven perspectives that help clients make informed, strategic decisions.

Embedding AI innovation across business lines is central to RBC Capital Markets’ vision of delivering unmatched client value, revenue growth, and competitive advantage. “We’re obsessed with our vision for the workflows,” says Bobby Grubert, Head of AI and Digital Innovation, RBC Capital Markets. “Everything we do is seen through the lens of solving client problems.”

This relentless pursuit of measurable, real-world AI outcomes prompted RBC Capital Markets to develop a proof of concept with NVIDIA and KX in 2024, building on an existing, trusted partnership. The goal? Augmenting analyst research with GenAI capabilities to expand coverage while simultaneously improving quality.

With 100 analysts covering 1,450 companies, scaling timely, data-led insights was a key driver of growth and competitive advantage for RBC Capital Markets. “Our investors appreciate great research; it drives client and commercial value,” agrees Bobby.

Leveraging its AI-powered genAI platform Aiden, RBC Capital Markets was determined to increase its coverage across companies and sectors.

Accelerated experimentation in a dedicated lab environment

Collaborating with NVIDIA and KX on a proof of concept in a dedicated lab environment accelerated RBC Capital Market’s AI experimentation by around four months. Within that lab environment, the team demonstrated significant performance gains:

• Up to 20X faster query and search speeds
• Up to 60% faster insight delivery to clients
• 20-second visualization times

KX’s unified platform for time-series and vector analytics enabled a potent blend of speed, scale, and reliability for these new AI capabilities. The shift from CPU-based pipelines to NVIDIA accelerated computing also fundamentally transformed RBC Capital Markets’s research workflows and the ability to surface new ideas. In the POC environment, our users were able to query millions of market and alternative datapoints with sub-second performance that’s up to 20x faster than RBC Capital Markets’s former solution for combined datasets.

Users could ask plain questions and receive fast, auditable, and clearly sourced answers based on millions of events, with real-time structured market data embedded directly into LLM responses.

Of course, such rapid and scalable insights mean nothing if they’re wrong. “Accuracy is critical to us, trust is critical to us, predictability is critical to us,” adds Bobby. RBC Capital Markets needed to know that the system could always be trusted to make informed, context-aware recommendations, necessitating robust guardrails and clear explainability.

From experimentation to production: Aiden QuickTakes

The insights from the KX and NVIDIA proof of concept directly informed the development of Aiden QuickTakes, a production capability that incorporates multiple specialized agents, including earnings and filings agents, that can shorten research times from hours to minutes.

Crucially, AI agents augment rather than replace human expertise at RBC Capital Markets, ensuring additional oversight. “Human in the loop is a very important philosophy for us,” emphasizes Bobby. “This is human research at AI scale.”

Armed with new AI capabilities, RBC Capital Market’s analysts have dramatically expanded ticker coverage, while also improving research quality. Analysts spend far less time gathering data and far more on generating strategic insights and engaging clients. For instance, the time to deliver earnings insights is now up to 60% faster.

By integrating AI-powered research assistance, RBC Capital Markets expanded its XL-ranked teams from 17 to 18 in both Canada and the US, securing six new analyst rankings, and strengthening its position as a top 10 research provider.

Deepening the partnership across front-office workflows

The collaboration with KX is now extending beyond research into front-office operations, with two initiatives currently in development:

• AI Email Agent: An AI-powered agent built over shared mailboxes that enables sales and trading teams to draft responses to client inquiries quickly and accurately. By integrating KX’s structured data capabilities with generative AI, the agent can surface relevant positions, trade history, and market context in real time, reducing response latency and improving the quality of client communication.
• KDB MCP Server: A structured data connectivity layer that serves KDB data directly to Aiden users, primarily in sales and trading. This gives front-office professionals natural language access to the firm’s time-series data, embedding KX’s analytics directly into the workflows where decisions are made.

Both initiatives reflect a deliberate shift: moving KX’s capabilities from experimentation into the operational layer of the business, where they can drive measurable impact across client-facing functions.

Looking ahead

As RBC Capital Markets continues its journey to becoming an AI-enabled investment bank, leveraging cutting-edge capabilities enabled by KX and NVIDIA will play a vital role. “We wouldn’t be able to do any of this without Borealis and our strategic partnerships,” concludes Bobby. “We’re prioritizing industry leaders that we go deep with in terms of strategy — NVIDIA and KX being two of those firms — as we go all in to scale up and out across capital markets.”

KX is proud to partner with RBC Capital Markets, delivering the unified data layer and high-performance analytics essential to realizing AI value at enterprise scale. Learn more about our partnership with NVIDIA and how we help firms drive real-world business outcomes with responsible, scalable AI.

Explore how KX and NVIDIA are scaling production-grade AI for capital markets.