/retail banking

News and resources on retail banking, consumer finance and reinventing customer experience in finance.

GTreasury unveils AI tool to improve cash forecasting

Source: GTreasury

GTreasury, a treasury and risk management platform provider, today announced the release of SmartPredictions™, which improves cash forecasting accuracy by applying artificial intelligence to short- and long-term treasury data.

As forecasting precision becomes increasingly critical to corporations’ cash management and liquidity planning, GTreasury is making the capability generally available to customers beginning today.

SmartPredictions uses AI to train and test business’ historical liquidity data in order to accurately predict and forecast future transactions. With an intuitive interface, treasurers can quickly indicate which of their datasets to include in a given forecast. They can also customize how much historical data to include, as well as how far into the future to request forecast data. SmartPredictions will algorithmically select the optimal model for each dataset input – based on predicted accuracy – from a variety of machine learning models as well as a traditional projection method that is well-suited for time-series data problems.

“Proactively planning for future liquidity requirements – with as much accuracy as possible – is more important than ever,” said Renaat Ver Eecke, CEO, GTreasury. “Corporate decisions hinge on cash forecasting and, as we’ve all seen this year, that forecasting can change rapidly. With SmartPredictions, we’re providing treasurers with an AI-fueled forecasting solution that gives them best-fit accuracy and will continually earn their trust. As treasurers are well aware, more accurate forecasting yields superior operational cash management, improved ROI rates, and more confidence in paying down debt.”

Two of the machine learning models that are at the core of GTreasury’s new SmartPredictions capability:
• Decision Tree Regressor (gradient boosted framework) builds out regression models in a tree structure – splitting historical data into increasingly smaller samples according to predictor variables such as day/month/year, early-month/end-of-month, and other seasonality indicators.
• Singular Spectrum Analysis processes data points sequentially to predict what subsequent data will look like based on previous influences.

“We built SmartPredictions to be transparent in how artificial intelligence is being applied to treasurers’ data,” said Roger Comins, Director – Product Management, GTreasury. “Our solution is uniquely clear about which forecast model was chosen as the most accurate fit, why it was chosen, and what the expected error margin is for every forecast. SmartPredictions offers treasurers the industry’s most accurate and least time-intensive cash forecasting solution.” 

Comments: (0)