AI Simulations Enabling More Accurate Forecasting

Andrea Matsuda at NICE explores how AI simulations are enabling more accurate forecasting.

The pandemic accelerated the adoption of digital channels for customer service, and the implications for contact centres extend far beyond capital expenditures and investments in technology to also include how they manage the contact centre’s most valuable asset—its workforce.

With more interactions taking place on digital channels, traditional phone-centric workforce management solutions have become increasingly limited in their effectiveness.

Contact centres find it hard to use them to forecast with any degree of certainty the number of staff needed at any given time, leaving leaders with little more than the hope that they’ll have the right number of people, with the right skills, on hand exactly when they’re needed.

But meeting the needs of today’s complex contact centre requires a strategy based on more than just hope. Modern workforce management solutions are increasingly turning to machine learning to accurately forecast and staff to handle multi-channel interactions.

Forecasting with Machine Learning

Machine learning is a subset of artificial intelligence (AI) that uses advanced algorithms that can learn from and make predictions based on data.

The machine learns by being “fed” substantial amounts of information (input data), and initial guesses are compared to the best answers or specified expected outcome (output data).

Machine learning generally occurs in one of two forms:

1. Supervised learning. The computer is provided with an example of inputs and the resulting outputs from those inputs, with the goal being to derive a rule or routine that processes the input to automatically predict the expected outputs.
2. Unsupervised learning. The computer finds structure or patterns in unadulterated input data with no prior knowledge of the resulting outputs. In this form, the goal of machine learning is to analyze the raw data to discover hidden patterns.

When it comes to forecasting, most legacy workforce management (WFM) systems rely on Erlang, a programming used to create scalable systems.

While Erlang has proven reliable in some areas, it fails to account for today’s multi-skilled, multi-channel worker, assuming instead that all employees share a homogenous skill assignment and that work items are tied to a single skill profile.

As a result, the system tends to overstate how many employees are needed per required line, leading to costly overstaffing at a time when margins are stretched thin.

Another limitation of Erlang-based forecasting is that it doesn’t factor in abandoned calls. To correct for this, some WFM systems attempt to apply an adjustment factor provided by the end user. Other systems try making an educated guess—a rudimentary estimation of the adjustment factor.

Modern workforce management solutions approach the problem another way. Using a form of supervised learning that is embedded within a simulator, our Workforce Management’s AI forecasting tool accounts for the performance of multi-skilled employees to produce a more reliable staffing estimate.

It automatically compares multiple forecasts to produce the best result—no statistics experts required. And, as underlying factors change, such as a shift in demand or the addition of multi-channel employees, the solution automatically runs simulations to figure out the impact on overall efficiency, further fine-tuning the process as it goes.

Our WFM also uses unsupervised machine learning to make skills-based assessments that are critical for contact centres looking to optimize scheduling and maximize the use of various digital channels.

Long-term planning also stands to have a significant impact on contact centres that invest in it. Enhanced Strategic Planner (ESP) is an optional module that helps contact centres capitalize on their full long-term planning potential by providing the answers and insights needed to make better workforce planning decisions.

By leveraging AI-generated forecasts that adapt to historical data, ESP enables more accurate predictions. With “what-if” planning scenarios for workforce and performance metrics, such as business acquisition or staffing changes, ESP helps contact centres not only make informed decisions about hiring but also react more quickly to business changes.

Ultimately, it all comes down to accuracy. Machine learning has shown to improve staffing forecast accuracy over older methods by 8%, and as AI technology evolves, this figure is expected to grow over time.

The days of contact centres “hoping” the best staffing plan is executed are coming to an end. Thanks to the power of AI, businesses will be able to predict how many employees are needed and on what channels, saving on costs and, hopefully, making for a better customer―and agent―experience.

This blog post has been re-published by kind permission of NICE – View the Original Article

For more information about NICE - visit the NICE Website

About NICE

NICE is a leading global enterprise software provider that enables organizations to improve customer experience and business results, ensure compliance and fight financial crime. Their mission is to help customers build and strengthen their reputation by uncovering customer insight, predicting human intent and taking the right action to improve their business.

Find out more about NICE

Call Centre Helper is not responsible for the content of these guest blog posts. The opinions expressed in this article are those of the author, and do not necessarily reflect those of Call Centre Helper.

Author: NICE