Vanguard’s AI investments pay dividends

Vanguard’s north star for AI has been “one of our biggest successes,” Wilkinson says, in advancing the company’s AI strategy from experimentation to business value.

“We set the tone from an early stage that AI is not a separate thing over here,” she says. “We set it up front that AI is in service of our mission to improve outcomes. It’s not something different.”

Early experimentation helped teams understand what generative AI could do, but success came when those learnings were paired with clear business priorities, defined measures of success, and strong governance, Manry explains.

Vanguard

“Senior business leaders play an active role in setting top‑down direction, ensuring AI is applied where it can drive real client and business outcomes, while teams on the ground continue to surface ideas through bottom‑up innovation,” she says. “Critically, each use case is anchored to a clear hypothesis and measurable outcome before it scales, making it easier to double down on what works and quickly stop what doesn’t.”

That north star vision also comes with the expectation that AI should deliver returns, Wilkinson says.

“When we make a bet, we better have measurable ROI associated with that, and we’re being disciplined about measuring it,” she says.

The result, according to Manry, is a disciplined approach that combines “strategy, process redesign, change management, and adoption to deliver measurable value.”

Creating culture change

One key to Vanguard’s success has been its approach to skill building — or, as Wilkinson calls it, the “enablement layer,” where both IT and business workers have opportunities to learn and experiment.

Early AI experiments at Vanguard happened in three areas: code generation, content creation and summarization, and search.

“These experiments were very internally focused initially because we were just trying to get our hands around not only how these tools work but also what needed to be in place from a responsible AI perspective,” Manry says.

Vanguard also empowered leaders and workers to determine where and how AI could deliver value.

To do that, Vanguard’s most senior executives, including global CIO Nitin Tandon, created AI champions — high-level stakeholders, including executives and senior managers from the business divisions with demonstrated AI literacy, who are “setting the business priorities for where to apply AI,” Manry says.

“We have representation across not only our client-facing businesses, but also our shared service functions — HR, legal, compliance,” she explains. “And those people are deputized with setting the top-down strategy for their divisions and thinking about what we want to do from an enterprise top-down perspective.”

The creation of AI champions has enabled Vanguard to move from “lots of experiments to getting more focused on where we can drive better client outcomes and better business outcomes,” Manry says.

Vanguard also appointed digital ambassadors. These are workers “right at the edge,” Manry explains, who are experimenting with ways to apply AI and identifying use cases where AI could drive innovation. This provides a bottoms-up perspective, “surfacing use cases that might not be part of the top-down strategy,” she notes.

Each business division performs its own prioritization of proposed AI initiatives. “It’s a very business-led activity,” Manry adds.

Picking winning AI use cases

Wilkinson credits Vanguard’s methodical tactics for picking winning use cases such as Expert Insights.

“We have been very strategic about where we’re placing our bets with AI,” she says. “All along we have viewed AI not as a tech or AI strategy, but as a business strategy. And in the FAS business, independent of AI, we had a goal to scale portfolio specialist access.”

The key for Vanguard has been starting with a business problem, rather than taking a tech-for-tech’s-sake approach.

“We had a business objective totally separate from AI,” she says, noting that in the case of FSA it was to scale the service more broadly. “We thought there was a great opportunity here [for AI].”

Bringing together a portfolio specialist and a methodology expert with an AI engineer and product manager, the FSA group was able to quickly move from a proof of concept, to a pilot, to a working solution — all while ensuring Expert Insights was delivering the right outputs.

“[Now we can] scale portfolio specialist analysis at a much higher rate than we’ve ever been able to before,” she says.