Eight years of data. One conversation away.
Their data held the answers, but analysis was manual. And took days. So Walker implemented a queryable AI agent. Now, they get answers in minutes.
This B2B2C wealth management firm manages over $500 billion in assets under administration. And supports a growing network of thousands of financial advisors. Understanding advisors’ needs—quickly—was critical. But analysis was performed manually. And with years of data in disconnected sources—fast was impossible.
No exports. No SQL to write. Just an answer.
The firm's entire survey history—every response, every source, going back to 2018—is now queryable in plain language. No exports. No SQL. No waiting on a data team.
Ask a question—get an answer.
Great data. Buried in many systems.
Five survey programs, measuring satisfaction at the customer, employee and transaction levels. More than 109,000 responses collected over eight years.
But each source lived on its own island: Qualtrics dashboards, third-party files, manual exports. There was no complete view.
The CX team had hunches about themes forming—a feature rollout confusing advisors, a service inconsistency creeping into every region. But confirming those themes using the data meant manual work. A lot of it.
The team read comments, one by one. Built reports from scratch. Assembling a cross-source answer could take days.
Every request, a long project. A small team, stretched thin.
Five survey programs. One place to ask questions.
Walker worked with the client team to understand where all of their data lived.
Then the work started. We brought eleven sources into one governed Databricks environment and moved eight years of responses with them. That meant five survey programs—relationship, transactional, employee, competitive and ongoing advisor feedback—plus three third-party data sets on advisors and their market, and three competitive intelligence feeds.
On top of the data, Walker built a custom Genie: a natural language interface that reasons across the data and returns analysis and visuals. No SQL to write. No manual digging.
The harder problem was who gets to see what. Survey data includes identifying details, and managers shouldn’t see how their own team responded to employee surveys. So we set up role-based permissions. External consultants see everything. Employees see aggregated results only. One dataset. Two views.
Survey responses—eight years of history, available for querying on day one
Survey programs unified into a single governed data cloud
Data sources connected
Don’t take our word for it.
“Being able to use AI technology through Databricks to streamline and make reporting more effective, easier and more timely—that was the main thing. A game changer.”
Hours of manual work—reduced to minutes.
The team tested the solution with a real need: understanding advisor sentiment about a recent multi-factor authentication tool rollout.
Answering that used to mean reading thousands of open-ended comments. Days of work. The team's focus shifted away from everything else.
Now, their analyst could ask Genie instead. It surfaced, tagged and summarized every relevant comment across all five survey programs. Reading any one source alone could have made the confusion look like a one-channel problem—or missed it entirely. Querying all five at once showed the team it wasn't isolated: advisors across channels were hitting the same friction.
The analysis—complete, across every source. And delivered in 15 minutes.
That's just the start. The next step is to connect advisor feedback with support call transcripts and CRM records. So Genie can tie what advisors say to what happens in the business.
Here’s a few key players.
See more of our work.
