For AI, it’s not Big Data, but Accurate Data that Counts

Having had a front row seat to data warehousing trends in the financial industry for the past twenty years, I feel somewhat qualified to state that I often cringe whenever I hear the term ‘big data’ thrown about recklessly, especially when it comes to AI. The truth is that to gain insight into your customers through behavioral science and predictive analytics, it is not ‘big data’ one usually requires, but meaningful and accurate data. This is why InvestCloud, with arguably the largest and most advanced financial data model on the planet, has placed as much focus on building data collecting and maintenance tools as the focus spent on the AI techniques we deploy. 

While InvestCloud’s AI algorithms are proprietary, we can disclose a few examples of how and why they are used. To begin, it is important to first understand the number one challenge in CRM applications on the market - their adoption rates barely reach 10%. There are many reasons for this low adoption number but the results are the same. Poor adoption = poor data collection = meaningless analytics.

 Before setting out to build our own CRM, InvestCloud realized we needed to address the issue of gathering and maintaining quality data. There is one CRM system on the market that is different from the others and that’s LinkedIn. In addition to their high level of product adoption, the accuracy of their data is through the roof because it is entered and maintained by the end user. An Advisor is not entering our LinkedIn data, we are. 

InvestCloud’s CRM, as well as our apps - Digital Onboarding, Digital Advice, Digital Financial Planning - have been built to gather data entered directly by the customer. This includes information about themselves such as their investment objectives, risk tolerance, KYC / account opening and funding information, and goals-based financial plans - which then becomes accurate and meaningful data that feeds our CRM. 

Data collection is only one half of data modeling. Accurately representing how information changes across time series is the other. A challenge with trying to represent historical information is that your data model must act like a time machine. In our industry, you have to represent a time series of data where there are multiple copies of the same period that reflect the cancellations, corrections and other changes that take place. Additionally, you need access to all of those copies. This capability is baked into our product and allows our clients to do high-end predictive analytics. 

InvestCloud has deployed a series of behavioral science, decision making and predictive analytic tools that help our clients grow their business and provide massive automation to help them service more customers. We set out to do this intelligently, recognizing that the real secret to success is to maximize user adoption, and thus collect as much information directly from the customer as possible.

Strong adoption + accurate data collection = cutting edge analytics.