Salesforce announced the pairing of its client relationship management system, Financial Services Cloud, and Einstein Analytics into an analytics solution for wealth managers, home office professionals and retail bankers, according to the firm. Called Einstein Analytics for Financial Services, the service uses Salesforce’s own query engine and interpretation layers, powered by the enterprise data analytics of its 2016 acquisition, BeyondCore.
The product comes with two prebuilt, but customizable, models, which measure client churn (or clients considered to be at risk of leaving a firm) and the likelihood a client might bring additional assets to a firm, as well as suggestions for improving those factors. The models are designed to measure the probability of a future scenario within the firm. An advisor interested in deciphering whether a client might leave could see that client’s characteristics against those of other clients in the firm, along with a readout of suggestions for improving the odds that a client will stay at the firm.
Home office professionals or data scientists interested in “seeing the math” behind the models can have inside access to them, and even make tweaks if necessary. While the tool provides benchmarking at an enterprise level, firms can input and run the model off of their own industrywide data, if they so choose, said Amruta Moktali, VP of product management for Einstein Analytics.
This isn’t the first time machine learning has been applied to the Financial Services Cloud. In 2017, Salesforce paired its financial services–focused customer relationship management with Einstein to help clients identify important life events and make suggestions for actions around those events. In 2018, Financial Services Cloud attempted to eliminate fragmentation between its consumer and commercial lines of business.
The impetus for providing a more wraparound product came from observing how customers were using Einstein data, noted Moktali. Customers were taking data and building their own dashboards or metrics, so the firm decided to prepackage what it was seeing across multiple users. “Our goal is to make it faster for our customers to adopt and get running and doing their job, rather than spending too much [time] trying to build stuff,” she explained. Even though there is plenty of room for customization with the new platform, which WealthManagement.com was able to preview, the goal was to “templatize” the product.
As is the case for other instances of machine learning and analytics, insights are partially determined by the quality of the data that’s ingested into the analytics engine. Firms using this tool will need to provide their own source of aggregated data, which then hits a normalizing filter before it’s brought into the Salesforce analytics engine.
“Data is data. The platform doesn't care where it's coming from,” explained Moktali. Missing data, however, could affect the quality of the insights. Some elements of the tool, like grouping individuals into households, still need to be manually inputted by advisors in certain cases.
More machine learning–based models are in the works, noted Umair Rauf, a senior director of product management at Salesforce. The next wave of models to be considered will look for ways to reduce heldaway assets and increase assets under management. The plan is to include those models and more, automatically, whenever updates are rolled out.
The tool was developed with feedback from a variety of sources, ranging from enterprise financial advisors down to firms with hundreds of advisors. “We at Salesforce don’t discriminate between a big enterprise company or a small company. We had to go across the board and talk to all sorts of wealth managers,” said Moktali. The service costs $150 per user per month and is not a stand-alone product, she added. It needs to be used in conjunction with the Financial Services Cloud or Einstein Analytics Plus.