×

A robot to explain your mutual fund statement

USAA prides itself on the personal relationships it's forged with the military families it serves. But with 11 million members who increasingly interface with the financial services company through digital means—it will soon count 2 billion digital interactions annually—maintaining that personal touch presents a significant challenge.

"We've got all of these sources of information," said Zacharay Gipson, USAA's chief innovation officer. "But how do we create personal interactions in our digital and mobile channels?"

Narrative Science
Source: Narrative Science

Seeking an answer to that question, USAA last year invested $10 million in Narrative Science, a Chicago-based software company and 2015 CNBC Disruptor, whose primary software product—named Quill—turns mountains of data into digestible written reports that read like they were penned by a human hand. In June, the two companies announced a deeper partnership, one that will see Quill deployed within USAA to generate personalized written reports for individual members' portfolios, seamlessly crunching dense piles of numerical data into easy-to-read performance summaries on demand.

For USAA, the partnership represents a next-generation solution for its personalization problem; members interfacing with USAA digitally won't just receive top-level data about overall fund performance, but personalized write-ups detailing the performance of their individual portfolios. For Narrative Science, it's the next logical step toward bringing high-powered data analytics down from the mountaintop and into the hands of every individual.

"The trend in the world right now is toward more and more personalized data," said Nick Beil, Narrative Science's chief operating officer. "But most people aren't data scientists. One day we'll look back at spreadsheets and we'll be amazed that we used to look at data that way."

From baseball to brokerage

Narrative Science began as a project at Northwestern University in which students and researchers built a program called StatsMonkey that automatically translated numerical baseball game statistics into crisply-written narrative game recaps. As the software improved, the write-ups became more natural, to the point that it became difficult to distinguish between StatsMonkey's recaps and those written by journalists.

In 2010 Narrative Science emerged from the StatsMonkey experiment as an independent, venture-backed company. Its software engineers used what they'd learned from crunching baseball stats to build a completely new artificial intelligence platform capable of turning numbers—any kind of numbers—into written reports.

Read MoreBillion-dollar startups you've never heard of

Called Quill, the software now has roughly 50 enterprise customers, more than half of which are in the financial services industry. Companies like T. Rowe Price, Credit Suisse, MasterCard and American Century Investments use the software in-house to automatically generate intra-day trading summaries for traders or fund performance analyses for portfolio managers. Clients in the intelligence community similarly use the software to quickly turn mountains of numerical data into digestible written reports. Forbes uses the software to rapidly crank out company earnings reports.

Quill's output, though, has largely faced inward, producing performance reports for wealth managers or executives rather than for individual clients. The USAA partnership represents Narrative Science's first foray into crunching data for the individual client and offers a glimpse into the future that Narrative Science envisions for its technology.

Deep personalization

"It's hard to get down to the granular level of the kind of information that Quill is pulling together for [USAA] because of the sheer scale of the organization," said Kris Hammond, Narrative Science's chief scientist. "There's an irony here, in that the the only way to get down to that deep personalization at the level of what's going on with a portfolio, of what's going on with your individual investments—the only way to get to that in terms of communication is to have the machine do it. So it's personalization through a mechanism that many people think of as impersonal."

That's a departure from what you see elsewhere in the industry, according to USAA's Gipson. For the most part, financial services companies looking to leverage their vast customer data into better personalization rely on single point solutions. A bank sees that a consumer is shopping for a car, for instance, so it sends the client an offer for an auto loan. "That's one action, one response," he said. "But when we talk about personalization, we really mean that we know and understand what's important to you and how we can serve you. And that involves a much larger breadth of data."

Read MoreLittle-known inventors who cashed in on Quirky

The USAA partnership and what it offers to the company's members is really just a starting place for a larger rollout of software like Quill to individuals, not just corporate clients.

"We think of Quill as the voice of the machine," Beil said. Both he and Hammond believe the software can be integrated into any kind of data analytic or visualization, from the plethora of health apps running on smartphones to the opaque pie chart that breaks down spending on some credit card statements. "You can discover things in data by seeing it in text," he said, "things you might not see in a chart."

The vast amount of data piling up in fields like health care, business intelligence, marketing, and education are low hanging fruit for the kind of insight Quill offers, allowing companies and even governments to better access their available data and make better decisions based on what they learn from coherent, text-based reports that take half a second to generate.

"We can free people from the need to be completely dependent upon data scientists or their own ability to understand data," Hammond said. "We can provide stories that are both readable and true, and that's a powerful moment."

By Clay Dillow, Special to CNBC.com