How TD Bank Is Building Trusted Enterprise AI with Strategy Mosaic's Universal Semantic Layer

Quick Answer

• TD Bank, one of the six largest banks in North America, is building its enterprise AI foundation on Strategy Mosaic's Universal Semantic Layer, deployed on Google Cloud.

• By governing business definitions independently of any single data platform, TD Bank's capital markets division is creating a trusted, scalable foundation for regulatory-grade AI across its trading and analytics ecosystem.

TD Securities, the capital markets arm of TD Bank Group, is the sixth-largest bank in North America by assets.

It operates across quantitative groups, trading desks, and electronic trading businesses, where specialists own their own data, models, and understanding of how the business operates.

That distributed expertise is a competitive advantage, enabling teams to develop deep domain knowledge and specialized perspectives. But without a shared semantic foundation, it can also cause enterprise AI to produce inconsistent answers at scale.

That's the challenge TD Bank is addressing by building a shared semantic foundation with Strategy Mosaic, ensuring every AI application starts from the same trusted business context.

As TD Bank accelerates its enterprise AI roadmap, its leadership is focused on scaling a governed data foundation across every team, tool, and AI workflow.

That's precisely why TD Securities is building an AI architecture designed to address data fragmentation before it compounds into larger governance and consistency challenges.

The approach is to deploy Strategy Mosaic's Universal Semantic Layer natively on Google Cloud as the governed foundation every tool, dashboard, and AI agent draws from, establishing shared business context before agentic workflows arrive, not after.

As Dan Bosman, SVP & Chief Information Officer at TD Securities says

"What's been really exciting is the bold vision around the Universal Semantic Layer. These two partners, Strategy Software and Google Cloud, are key to our success. The platforms complement each other really well."

To bring that governed foundation to life, TD Securities is deploying Strategy Mosaic as an independent semantic layer that sits above the underlying data sources and below every consuming tool.

Instead of being tied to a single vendor's compute engine, the semantic layer is owned and governed entirely by the enterprise. That distinction is especially important in capital markets, where regulated institutions can't afford to:

• Have critical business definitions locked inside a single platform

• Expose conflicting metric definitions to regulatory scrutiny

• Depend on a single vendor's availability, pricing, or product roadmap

An independent semantic layer addresses these risks by decoupling core business logic from vendor-specific limitations and establishing a single, governed source of truth across the enterprise.

As a result, whether someone queries data through a spreadsheet, a BI dashboard, or a custom AI application, they receive the same governed business answer.

For TD Bank, that means critical business definitions such as VWAP (volume-weighted average price) are governed once within Strategy Mosaic and made consistently available to every tool and AI agent across the enterprise.

Dan Bosman described the moment the architectural vision clicked. During his first Strategy World event, he saw an early demonstration of Mosaic:

"The light bulb went off in my head. I saw the value of this semantic layer, and I saw a need for a pivot; to do something different than the next data platform, the next data warehouse being the winner."

By establishing a governed semantic foundation, Strategy Mosaic is designed to help TD Bank scale agentic AI within a consistent, trusted environment without sacrificing the autonomy individual teams need to innovate.

Because AI models consume data shaped by consistent business logic through Strategy Mosaic, they can surface insights faster while producing more reliable outputs across complex capital markets workflows. At the same time, individual business units can continue developing specialized analytics and strategies while operating from the same governed semantic context.

As a result, TD Bank is positioned to scale AI initiatives across strategic growth areas such as:

• Transaction banking

• Electronic trading

• Client-facing analytics

For TD leadership, the objective extends beyond technology. As Dan Bosman explains, AI should strengthen human expertise rather than replace it:

"We are a digital-first organization, but we are proudly remarkably human. If we can use AI to take the toil away and refocus our efforts on what is important — our clients, and these human connections — that is what freedom means."

By combining trusted business definitions with governed data access, TD Bank is building the architecture needed to scale enterprise AI while maintaining the consistency, transparency, and trust that a highly regulated environment demands.

With Strategy Mosaic serving as its Universal Semantic Layer, the bank is laying the foundation for AI that can scale without compromising governance or business context.

For enterprises pursuing the same goal, the challenge isn't simply scaling AI. It's ensuring every model, application, and team operates from the same trusted understanding of business context. That's the architectural approach TD Bank is taking today.