Trace Id is missing
February 27, 2020

PureFacts extracts financial insights in seconds and calculates fees more accurately with AI

PureFacts helps wealth management firms retain investors, manage risk, maintain compliance, and run operations better and faster with AI-powered solutions—empowering its customers to provide more personalized experiences for their investors at scale. The company built its “Insights as a Service” platform on Microsoft Azure. It uses machine learning to reduce client attrition, extract financial insights in seconds, and catch 80 percent of fee billing data anomalies with 90 percent accuracy.

PureFacts

“Tools like Azure Synapse are a fantastic way of leveraging big data plus AI. Layering Microsoft managed services on top of that really gives us the ability to scale more than ever before.”

Steve D’Costa, Chief Technology Officer, PureFacts

PureFacts has attracted many customers in its 10 years in business based on the accuracy of its fee-calculation solutions alone—companies that range from boutique investment firms to wealth divisions of large banks and insurance companies. Considering that a fee calculation error can result in a fine from a regulatory agency and plenty of bad publicity, this is quite a vote of confidence.

PureFacts employees take this trust seriously, and they enjoy what they do. “We’re data geeks and wealth nerds. Our water-cooler talk is finance, and we do math equations on the whiteboards for fun,” says Andrea Boileau, Director of Human Resources at PureFacts.

An evolving market requires a new approach with AI

As the 2020s begin, the industry faces an unusually large transfer of wealth in the next 10 to 15 years. PureFacts incorporated AI into its “Insights as a Service” platform to provide the hyper-personalized wealth advice at scale that a new generation of investors expects. Designed to help customers retain investors, extract decision-critical insights, and provide data anomaly detection, the platform uses AI and machine learning to automate and oversee data analysis and improve solution accuracy over time, to help lower risk and upfront investment.

Investor retention model to improve service, gain a competitive edge

For example, PureFacts’ investor retention model predicts whether an advisor’s investors are likely to leave in 3, 6, 12, or 24 months, based on more than 150 variables. Advisors see a probability assigned to each possible outcome and insights into the factors driving each one. With “what-if” modeling tools, they examine the top-two drivers for each prediction and estimate how various changes in their approach will affect the investor’s probability of leaving.

Talking to data: an intuitive way to extract key insights

PureFacts customers also use the platform to talk to their data and extract insights quickly. In the past, investors might call and ask what kind of risk exposure they have in Asia. Advisors had to sign into multiple systems and pull reports on all the investors’ accounts with exposure in Asia, scanning just the relevant parts of each mutual fund. It could take 30 minutes to an hour to collect all the information and formulate an answer.

With the PureFacts platform, advisors can query an investor’s portfolio allocation in Asia and get a report in seconds. The solution’s natural language processing model interprets advisors’ questions as SQL queries.

Data anomaly detection to reduce risk

With machine learning, advisors proactively identify and correct anomalous data patterns faster. It used to take a few days to correct the errors from mis-valued or mis-priced assets from custodial or portfolio-management data, and during that time, advisors couldn’t process fees or issue account statements.

The PureFacts platform uses machine learning to identify anomalous patterns in investors’ data and flag them for financial advisors. It then captures advisors’ feedback on how they dealt with that anomaly, such as whether they confirm everything is normal or continue to investigate. PureFacts uses machine learning on this data to retrain the model to become more accurate.

Ultimately, says Victor Skrylev, Vice President of Product at PureFacts, the company will have enough data and model training to let the solution start fixing anomalies automatically. “Our approach is different from the common method of defining and building risk controls around certain anomalies that solution providers are already aware of,” says Skrylev. “It’s easy to miss the rare occasions when new systemic issues arise, and it means you’re not really providing a proactive mechanism for finding issues and ultimately preventing costly mistakes.”

Fee calculator that improves efficiency and mitigates risk

While PureFacts is investing heavily in AI-powered solutions, its original core platform for fee-based wealth management firms has become an industry standard in Canada and is ready for international expansion. Customers, including leading banks, use the company’s fee calculator technology to improve operational efficiency and mitigate risk. These institutions must process fees accurately to comply with regulatory bodies’ requirements around fee amounts and reporting. This may sound straightforward, but banking and investment fees are based on factors like the type of account, how long funds have stayed in that account, how the amounts changed during a given period, how well investments performed, and which monies belong to which investor across accounts owned by families.

The company also has a solution to make standard account reports and statements clear and transparent, delighting its customers and their investors with a time-saving, straightforward view into investments and fees.

Development flexibility and solution scalability with Microsoft Azure         

For its fee-calculator solution, PureFacts developers use .NET Framework and .NET Core services, controlling .NET streams through Azure Event Hub, which they set up to manage messages and queues inside the system. The company hosts customers’ solutions on-premises and in the cloud, and it uses Microsoft SQL Server.  

With Azure Machine Learning, developers have the freedom to use their favorite open-source frameworks and languages. The development team is also enthusiastic about the no-code designer and Azure Cognitive Services.

Cognitive Services is a set of APIs, software development kits, and services to help developers build intelligent applications faster—apps that can see, hear, speak, understand, and even begin to reason. Steve D’Costa, Chief Technology Officer at PureFacts, adds, “With the no-code designer, we accelerate model creation with the automated machine learning UI and access built-in feature engineering, algorithm selection, and hyperparameter sweeping to develop highly accurate models.”

PureFacts also uses Azure Functions with Kubernetes-based Event Driven Autoscaling (KEDA) to scale its solutions, plus Azure Cosmos DB for single-digit-millisecond data access to APIs, like SQL. PureFacts moves some customers’ on-premises systems to Azure with minimal coding thanks to Azure Functions and managed instances of SQL Server Database Engine, for automated database management.

Scalability is key because PureFacts solutions capture several terabytes of financial data. To make accurate predictions and glean insights, developers need this vast quantity of information to build a data model and retrain it as needed.

“Before Microsoft Azure, we were hitting the limit on how much data we could handle on-premises,” says D’Costa. “Now, we put terabytes of data into Azure SQL Database or Azure Synapse for enterprise data warehousing and analytics, then run Azure Databricks or machine learning tools on it without worrying about infrastructure.” He adds that, thanks to the underlying base code, PureFacts can train models and conduct fail-fast development more quickly and efficiently.

D’Costa adds, “Tools like Azure Synapse are a fantastic way of leveraging big data plus AI. Layering Microsoft managed services on top of that really gives us the ability to scale more than ever before.”

Benefits of partnering with Microsoft

A Microsoft Partner Network member with Gold Competencies in Cloud Platform deployments, PureFacts has a well-established rapport with Microsoft. Microsoft AI experts have met with PureFacts developers and other team members on-site and at Microsoft campuses to provide tutorials and other product expertise.

Skrylev says, “The knowledge Microsoft has provided helps us drive our AI strategy forward and saves us a lot of time and costs in trying to learn everything on our own.” For Skrylev, seeing Microsoft push the AI boundary with its products, and recognizing the development time his company saves, demonstrates that PureFacts has found a valuable technology partner.

PureFacts also gets sales support from Microsoft, bringing Microsoft representatives to meetings with several potential clients. Skrylev says, “Microsoft has a lot of relationships with bigger banks and other clients in financial services. So being able to say that we host our solutions in Azure—and having Microsoft representatives there to answer product questions—gives these companies a lot of confidence because they view Azure and Microsoft as trusted partners.”  

Benefits for PureFacts customers

Wealth management firms that choose the PureFacts platform can expect significant benefits, ranging from business growth to more efficient operations and reduced risk. These companies typically have lean teams—in some cases, just one person dedicated to processing fees every quarter. The task usually takes about eight business days to do, including handling any issues that might arise. With the PureFacts fee calculation solution, these firms could potentially reduce this time to two or three days—three to four times faster.

Retaining up to $120 million in saved business

When it comes to the business value of the investor retention technology in the PureFacts platform, Skrylev cites the average lifetime investor value at about $120,000. If a company reduces its turnover from 3 percent to 2 percent on a roster of 100,000 investors, it adds up to $120 million in saved business.

Extracting insights in seconds

Advisors who once took half an hour to extract insights for investors can type queries into the PureFacts platform instead and get the answers in seconds.

Confidently detecting more anomalies

With data anomaly detection, many banks run about four different types of reports on the quality of their data and then look through the reports line by line to find anomalies. “Our data-anomaly model today can catch 80 percent of issues with 90 percent confidence,” says Skrylev. “That’s potentially cutting an incredible amount of work, while identifying anomalies with far more confidence versus using manual methods.”

Find out more about PureFacts on Twitter, Facebook, LinkedIn, and YouTube.

“We put terabytes of data into Azure SQL or Azure Synapse … then run Azure Databricks or machine learning tools on it without worrying about infrastructure.”

Steve D’Costa, Chief Technology Officer, PureFacts

Take the next step

Fuel innovation with Microsoft

Talk to an expert about custom solutions

Let us help you create customized solutions and achieve your unique business goals.

Drive results with proven solutions

Achieve more with the products and solutions that helped our customers reach their goals.

Follow Microsoft