Recognos Details How AI Can Help Back-Office Processes

As artificial intelligence (AI) techniques improve, it's only natural that they will seep into newer and more varied functions throughout the capital markets ecosystem.

For the March issue of Waters, I detailed how machine learning is being deployed for front-office trading operations, specifically for market-timing strategies and other predictive processes.

Machine learning can be used to help improve risk management and surveillance systems. An article in this month's issue of Inside Reference Data, a sibling publication of Waters, looked at how machine learning is being used to improve data quality.

Recognos Financial has deployed machine learning, natural-language processing and semantic technologies for its Reference Data Center (RDC), an industry utility that serves as a central repository for documents relating to all US-based open-ended mutual funds. Yesterday, the New York-based vendor announced enhancements to the service, including a new alert system for when changes are made to Securities and Exchange Commission (SEC) documents for an open-ended mutual fund, as well as new analytics capabilities.

Drew Warren, the company's CEO, says the vendor uses AI technology to generate reference data. "Reference data is derived from complex documents," Warren tells Waters. "Most vendors have a small army of clerical staff—usually offshore—who read through a document searching for specific attributes. They then key this data into their system. We use the technology to do this function for us and we don't key data. ... Artificial intelligence can help users to gain actionable information in a number of ways. We focus on structuring unstructured or text data so that it can be used in a firm's big data initiative or in more targeted ways."

On the Horizon

For the company's latest release, automation was the key, says Ira Cohen, executive vice president of asset management services for Recognos. Historically, intermediaries have had to rely on humans at the fund to fax or email changes to the fund's policies and procedures. "Timeliness and accuracy of that communication has always been debatable," he says.

Cohen says the new RDC alert feature does not rely on a fund to communicate these events. "An intermediary can identify its funds and its specific data-point interests, and immediately upon a fund filing an e-mail will be sent to the intermediary highlighting the change along with the corresponding document," he says.

The other piece of the upgrade will help to make the data provided by the RDC database "more consumable" for the subscriber, he says, where the user can download the data and graphs into their existing databases and internal applications.

Looking ahead to new capabilities that will be rolled out over the next few months, Warren says they are currently integrating the Recognos platform with both business intelligence platform provider Smartlogic and its MarkLogic semantic layer. "This combination will provide enormous capabilities for firms, particularly in any big data initiative," he says.