In Episode 2 of Enterprise Aligned AI, I sat down with Vinh Tran, VP of Data and AI Platforms and RBC Fellow at Royal Bank of Canada, to discuss one of the most ambitious AI journeys in global finance.
While many organizations are stuck in the “pilot” phase, RBC has made a public commitment to drive $700M to $1B in value through AI by 2027. The key insight is the platform they have built to scale their human transformation. RBC is delivering on nine elite AI projects and also lifting every employee in the bank with AI.
“We’re driving the value not just from the six or seven or eight thousand developers or data scientists, which is historically where AI was being utilized. We’re now looking at transforming the 96,000… everybody with a computer: every mortgage broker, every engineer, every branch manager.” — Vinh Tran
The Power of “Self-Service” Agents
RBC’s core strategy is to move beyond a central “experts-only” model. They have built a self-service platform, headlined by RBC Assist, that empowers everyone from mortgage brokers to branch managers to build their own agents without writing code. What this unlocks goes beyond productivity.
“We have an opportunity to change how everybody works, how everybody does their job, drive more efficiency and productivity, but also be able to do things that people were never able to do before.” — Vinh Tran
Vinh notes the platform tools allow staff to explore complex data across multiple sources using natural language. In doing so, RBC is turning their workforce into a massive engine of builders.
Standardizing for Security, Reliability and Speed
To make “AI for All” a reality in regulated industries is no small feat. RBC focused on standardizing the “connective tissue” of their stack:
- The Data Foundation: This transformation is fueled by Lumina, RBC’s modern data platform. Vinh credits his predecessor for starting to build a robust semantic layer. The organization is now adopting a data-product mindset with curated data hubs (retail credit, mortgage) that give agents the semantic layer and high-quality metadata they need to understand banking fields.
“A lot of that work was underway well before our agentic journey. Now that we’re moving quickly on this agentic journey, we’re seeing that that semantic layer, that data classification, that metadata is very important to the agents.” — Vinh Tran
- MCP Gateway: RBC uses the Model Context Protocol (MCP) to standardize how agents access tools and data. By placing MCP servers behind a secure gateway, they ensure every agentic action is approved, compliant, and has a “human in the loop” for material system changes.
“We’ve put our MCP servers behind a gateway. That gateway has metadata on the tools and the MCP servers behind it, so we know that if you’re calling this tool, you better have some human in the loop and a plan that’s approved before you can do that.” — Vinh Tran
- Observability with OTEL: Standardization on OpenTelemetry (OTel) ensures continuous evaluation of agent plans and actions from production telemetry, not weekly batch runs.
Build vs. Buy Calculus has Shifted
It’s a better time than ever to build, and Vinh is passionate about it. He argued that three shifts have lowered the bar to build inside enterprises: open source that’s now unambiguously production-grade (the old “you can’t run open source in production” fear is dead), new AI coding tools are giving engineers 10x to 100x leverage, and cloud services that used to be the domain of large vendors are now deployable by anyone. You don’t have to write every line of code to build, you can be an integrator composing with what’s already out there.
In regulated enterprises, building is of even greater importance. When compliance and policy enforcement are non-negotiable, building provides greater control over policy guardrails and timelines. Vinh also argued that building is a culture-forging function for the engineering org, something you can’t buy.
“You’re generally going to save money by building, but I don’t think that’s the most important thing. When you’re in a regulated industry like we are, having control of your destiny, being able to incorporate your policies, your guidelines, patch your code, do everything that we need to do to meet regulatory requirements, it’s powerful for us to be able to have that control in our own hands.” — Vinh Tran
From Engineers to “Builders”
Vinh described the elevation of the engineering role. In this new world, engineers are no longer just writing code; they are becoming Uber Builders who orchestrate AI tools, audit model outputs, and ensure compliance.
“I think we’ll still need engineers, and I think they’ll become builders… the journey right now is what’s important: building those muscles, knowing how to use those tools.” — Vinh Tran
RBC’s $1B transformation proves that the most powerful accelerator in enterprise AI isn’t the model you buy. It’s the platform you build to empower your people. My favorite quote from Vinh in this episode:
“I really feel that we have to have the courage to build. We’re not going to write every line of code. We’re going to use the tools. We’re going to use the cloud. We’re going to use the community, but I think that allows us to really drive benefits for the organization and our clients.” — Vinh Tran
📖 Full episode with timestamps, key takeaways, and show notes
