I have heard about artificial intelligence (AI) being used in workforce management (WFM) solutions. How much historical data is needed for AI to improve forecast accuracy?


Not surprisingly, the more data you have the more accurate the findings. AI-enabled WFM solutions leverage machine learning (ML), an AI technology that is effective at finding patterns. ML is being used to identify outliers or deviations when validating models and forecasts in an iterative learning process, as well as to automatically identify the algorithm best suited for each set of forecasting criteria. A WFM solution fires off a number of algorithms to address a set of data and the ML engine identifies which model, from among as many as 45 different algorithms, works best with a given set of historical data. Predictive modeling, another AI-enabled capability, is used in forecasting and long-term planning to evaluate comprehensive “what if” scenarios.

That’s the easy part of the answer. If you’re running a small contact center on-premise, it could take you years to compile a data set large enough to provide optimal results. However, if you’re using a cloud-based solution and the vendor is able to combine your data on a randomized basis with many other cloud customers (also on a randomized basis), you could have access to the volume of data you need right away.