Northern Trust launches machine learning feature to aid asset allocators

New feature leverages users’ past behaviour to tag research management documents, allowing for further automation within the document management workflow.

By Wesley Bray

Northern Trust has launched a feature within its Front Office Solutions platform which utilises machine learning to provide better oversight of asset allocators’ research processes.

The new feature aligns with Northern Trust’s initiative to bring greater efficiency and transparency to alternatives asset servicing.

On a monthly basis, asset allocators receive thousands of files which are manually filled for investment manager research purposes. Northern Trust’s new feature is able to leverage knowledge of a user’s past behaviour to tag research management documents with suggested attributes before they are added to the Front Office Solutions platform.

Alternative investments such as private equity, real estate and infrastructure have typically required complex, manual processing. However, with the introduction of technologies such as artificial intelligence, blockchain and cloud technology, alternative investments have seen a higher level of automation.

Document procurement and document digitisation solutions have also been announced by Northern Trust, with the aim of providing more efficiency and improved oversight for asset owners investing in alternatives.

“Front Office Solutions embraces innovative technologies like machine learning to make investment teams more productive by providing faster access to the information they need to make portfolio decisions,” said Melanie Pickett, head of Northern Trust Front Office Solutions.

“The new tagging feature is an important step toward our ultimate goal of a fully automated document management workflow for alternative assets, which will improve efficiency and transparency.”

Due to allocators having different needs, the new tool is entirely tailored by the asset allocator.

Categories used to tag documents are created based on historical activity and the machine learning model continues to refine suggestions based on user’s interaction with the feature. Learning occurs in real time.

“Machine learning promotes significant efficiencies to the administrative work that goes into portfolio management, reducing staff time and improving research results,” added Pickett.

“With the administrative burden lightened and the possibility of human error limited, complex asset owners can focus on making better allocation and trading decisions, increasing the likelihood of improved operational alpha.”