The answer: Vanguard launched Expert Insights, a new artificial intelligence tool that takes a client’s holdings and produces clear, personalized investment analysis that an advisor can act on immediately. Portfolio analysis engagements have nearly quadrupled over six years, the company said, and the tool is designed to meet that demand without growing headcount.

Expert Insights is currently in pilot with select advisors and will be rolled out more broadly later in 2026. It’s the latest in a growing suite of AI tools Vanguard has built for advisors, including software that summarizes market commentary for clients and a stress-testing feature that models how a portfolio would hold up in a downturn.

Before the Tool, the Plumbing

Building Expert Insights required solving a problem that had nothing to do with AI. The firm’s analysts had always struggled to get basic data questions answered quickly. A simple query could take days, routed through specialists before anyone got a response. When Vanguard started building AI tools, it discovered those delays weren’t a technology problem. They were a data problem. The information existed, but it was scattered, inconsistently defined and not organized in a way any AI system could reliably use.

Fixing that meant doing work most firms don’t talk about publicly. Different teams had to agree on common definitions like what counts as a customer, how a key metric gets calculated and who owns which data. Compliance and security teams had to be part of the build from the start, not brought in at the end. The firm also built a library of more than 50 examples showing how plain-language business questions should translate into database queries, so the AI would produce answers in Vanguard’s own terms.

BlackRock Is Running the Same Race

Vanguard isn’t alone in discovering that the hard part of AI adoption is organizational, not technical. BlackRock, which manages some $14 trillion in assets, has been at this longer and is further along. Its experience shows both what’s possible and how much remains unsettled.

According to The Wall Street Journal, the firm recently rolled out RockAI, an internal platform that lets employees build AI tools for their own work in minutes, without any technical background. It launched to 5,000 software engineers last week.

BlackRock has already put AI to work in specific areas. A research platform called Asimov monitors investment theses around the clock, reading earnings reports and regulatory filings so that roughly 150 equity analysts don’t have to. A separate system helps client-facing staff prepare for meetings by pulling together relevant information before they walk in the room. The firm is now building an HR tool to fill out and submit paperwork on an employee’s behalf.

Nish Ajitsaria, who leads the effort, told the Journal that the end state he’s working toward is one where AI handles most routine processes and humans focus on judgment and oversight. “We can talk about reimagined workflows,” he said, “but that’s an implementation transformation initiative that doesn’t just happen overnight. I don’t know that anyone has cracked the nut.”

The Part That Can’t Be Rushed

The business case behind both efforts is the same. Firms that can answer a client’s question in seconds rather than days can serve more clients without growing headcount. Vanguard’s near quadrupling of advisor engagement is evidence the demand is there.

But the Vanguard and BlackRock experiences point to the same constraint: AI tools are only as good as the data underneath them. Firms that spent years organizing and governing their data find the groundwork done. Those that didn’t are doing it now, at the same time they’re trying to build.

For Vanguard, that meant getting engineers and compliance officers to agree on definitions before a single AI tool was built. For BlackRock, it meant nearly a decade of internal AI investment before RockAI was possible.