Originally published by Clara Shih on LinkedIn: From Regulatory Compliance to Relationship Insights: Why Financial Services is Uniquely Positioned to Benefit From the Data Analytics Revolution

In the early Hearsay years, our customers came to us because they realized social media was here to stay and, to take advantage of it, they needed to satisfy the complex record-keeping regulations that govern advisor-client communications in financial services. Little attention was paid to these records, aside from the need to retain and review them for compliance infractions.

Eight years later, with tens of millions of client contact records and interaction activities logged across Facebook, Twitter, text messages, emails, and website visits, we’re now embarking on a journey with our customers to do much more – harness this data to drive powerful relationship insights for the 150,000 advisors on our platform.

According to McKinsey Global Institute, the application of big data analytics to healthcare and government alone could yield $300 billion in savings and value creation in the next four years. With over $60 trillion in global wealth assets under management and $4 trillion in insurance premiums written last year, the wealth management and insurance industries could stand to gain even more. There are a few reasons for this.

First, the financial services industry benefits from having complete and reliable data, thanks to a clear compliance mandate from industry record-keeping and supervision rules such as SEC Rule 17a-4, FINRA Rule 3010, and FINRA Regulatory Notices 17-18, 11-39, and 10-06. Data analytics are only as good as the data that goes in, as summarized in this recent Capgemini paper.

Second, having a human in-the-loop, such as a wealth advisor or insurance agent, results in the greatest efficiency and accuracy in decision-making. Firms like Morgan Stanley are investing in predictive technologies that serve up prescriptive next-best actions for their 16,000 advisors based on data from Hearsay and other sources.

Finally, as Andrew Ng and others have proven, the most accurate machine learning is domain-specific, since you can define constraints and make assumptions that focuses the algorithm. There are only so many topics, for example, that one ever discusses with her insurance agent. Specifying that constraint narrows down the analytical problem to one that’s much easier to crunch. Venture capitalist Michael Yamnitsky has a nice explanation of this in his TechCrunch op-ed:

“Narrow data sets to shorten training periods and get to market faster. AI systems thrive off brute statistical analysis on big data. But much like teenagers, they need a little help being pointed in the right direction. The cloud business app is a narrow enough domain for the fledgling AI to flourish. Domain-specific data acts like the high-school basketball coach, helping the untrained AI identify relationships between inputs and desired outcomes, and shortening the training period necessary to run with success.”