Voice of the Client: How Experian is using AI to transform the business

By | 3 minute read | January 30, 2019

As one of the three major credit reporting agencies, Experian collects tons of data on individuals and businesses. From 3,000 sources of data we update 200 million records a month on that data.

Our U.S. Business Information Services unit (BIS) has been around since 1978, so we like to say that we’ve been doing data in a big way since long before size mattered so much.

Experian is also well-known for its data and analytics research and development. It maintains data labs all over the world staffed with mathematics wizards, data scientists and engineers — all of the key ingredients to help us along our journey to AI. But as talented and prepared as our quants are, they’re dedicated to real-world client problems and making our products scream with value. Within the commercial business unit, where U.S. businesses are our primary subject, it was early 2007 when BIS last considered a significant change to the systems infrastructure where we store and organize our primary data assets.

At the start of 2018, we decided this was the area where we needed to explore using AI for our own transformation.

Every company has its own set of problems that it attempts to solve. In our case, we needed a more efficient and accurate way to identify the relationships between businesses on which we maintain data. Typically, we’d supplement third-party verified data with manually compiled research on business families, corporate hierarchies and legal structures, meticulously linking these complex and volatile relationships.

Challenging? Yes, especially given today’s dynamic business environment where new acquisitions and divestitures occur regularly.

Even with knowledgeable, well-trained agents, publicly available data and specialized tools, keeping up with this dynamic business environment, global reach and ever-changing affiliations was outpacing the massive scale and high cost of our operations. To keep pace, we had to focus efforts on the top 5,000 global multinational corporations.

We captured real-time news events and data feeds to inform our agents and establish alerts, but still the business playbook — especially the critical information about which companies owned others, who was acquired by whom, who’s calling the shots and who’s executing the plays — was impossible to know with extreme accuracy for all but the big-league players.

For our clients who are interested in subject matter experts and need to move fast, we needed a better solution that could harvest data from anywhere, combine it with what we already know from our extensive data sets, churn through the massive computations and comparisons and return a simple, fast answer to the key question: “Are these businesses related?”

Enter the IBM Data Science Elite team

That’s where AI would be our natural problem solver, but again, this was something in which we had little expertise. To tackle a problem this big, we approached the IBM Data Science Elite team, who we knew could guide us through each waypoint on the journey.

Thanks to these stellar IBM resources, we had dedicated coaching from data scientists who not only shaped a proof-of-concept out of the dozens of possibilities in our environment, but also jumped in as player-coaches themselves, determining analytic approaches, designing and coding algorithms, validating, and measuring results – all the while mentoring some of the junior players on our team. What I truly appreciated then, and now, is the tremendous curiosity and creativity that they have infused into our environment and the high standards to which they hold themselves and all of us.

Get a look at the Data Science Elite team’s tools and techniques behind the Experian Proof of Concept from Carlo Appugliese in this companion blog.

A complete proof-of-concept with results

Today, AI and machine learning are helping to solve our problem with building and maintaining business families and corporate linkages with a potential 500 percent increase in coverage and 80 percent reduction in cost. Importantly, we learned that we have only scratched the surface of possibilities in finding business relationships of many kinds, and this will serve to drive competitive advantage for many years to come.

Learn more about how to succeed on your AI journey.