Workforce AI: The Driverless Contact Center


Are there any two words that strike more hopeful excitement and visceral fear into the hearts of men and women than “Artificial Intelligence”?  We revel in the prospect of robots and software that can shoulder difficult tasks for organizations and human beings as individuals, while being fearful that robots could get smarter than we are as human beings.

Although there’s some exaggeration and science fiction in the mix, it’s far more than just hype.  Artificial Intelligence (or more commonly just “AI”) is already here.  It’s in the iRobot that cleans your floors.  It’s the intelligence in driverless cars.  It’s the algorithms that filter spam out of your mailbox.  It’s in the Uber app that minimizes your wait time after you hail a car.  It’s in the autopilot features of commercial airlines that do most of the work in guiding a aircraft from one location to another.

According to the European Union, the AI market is expected to grow from about $664 million in 2017 to $3.88 billion in 2025.  That’s a jaw-dropping compounded annual growth rate of about 25%, and you can’t find many industries with that kind of sustained growth expectations.  The AI market is somewhat unique, as far as new technologies go, because we have a clear standard for what the technology should be able to achieve: the same physical and intellectual capacity that we have as human beings.  With such a clear demarcation of success, you can bet that scientists, mathematicians and many technology companies will be advancing the technology until we do achieve that goal.  We will leave predictions of subsequent advances to the science fiction movies.

In this era of much talk about AI, many vendors are using pure marketing to wrap themselves in the warm blanket of AI without giving much detail or evidence on what is actually possible with their AI-enhanced products.  Looking specifically at the contact center industry, we have seen a recent flurry of activity around chatbots, intelligent IVRs and virtual agents.  In a world where the customer experience is the holy grail, contact centers naturally apply new technologies first to their customers and then to their agents and internal processes.  Vendors are now starting to look at how AI can be applied to enhance the performance of agents, supervisors, managers, workforce planners, quality analysts and other members of the contact center workforce.  At Aspect, we call this new application of Artificial Intelligence “Workforce AI”, and it portends radical change in the way contact centers will operate.

AI will not take the contact center workforce by storm but rather in a gradual, carefully scrutinized evolution of capabilities.  In fact, the process of converting the management of a contact center to an automated environment, is likely to be much like the process of perfecting the AI being used in driverless vehicles.  Why?  Because there’s too much to lose if the AI does not work properly.  In the case of driverless vehicles, bad decisions result in injury and death.  In the case of the contact center, bad decisions result in unhappy customers, damage to the brand, loss of profit, poor employee morale, and lawsuits, among others.

How are driverless cars being created and tested?  Google’s Waymo uses a mix of driving simulations and real-world experiences to teach the driverless car AI.  Basically, they start with simple road challenges, but they keep a human driver in the car to take over if the automation does not know what decision to make or makes a bad decision.  The knowledge about the proper on-road decision made by the human driver is then incorporated into the AI of the driverless car.

We will see a somewhat similar process being used as the contact center workforce operations become increasingly automated.  We will start with simple workforce activities and decisions, then gradually move to more and more complex actions as the Workforce AI becomes increasingly trusted to make the right decisions, with humans monitoring these decisions and ready to override the workforce AI if it goes astray.

One of the beauties of artificial intelligence is that the type of AI technology used to simulate human intelligence is independent of the information sources used to make decisions and the actions that can be taken based on these decisions.  Suppose we wanted to design an AI solution to send out broadcast messages to agents regarding the availability of voluntary overtime.  We would want access to the same information that a human would use to make this decision, e.g., queue depth, number of agents currently working, skill types of available agents, history of changes in call volumes under similar circumstances, etc.  We could use a simple Boolean equation or a decision tree or a deep neural network to make the decision.  The point is that we can and should evolve the sophistication of our Workforce AI as the technology proves itself to be increasingly available and successful.  We will evolve from one type of AI to the next yielding increasingly “intelligent” decisions and associated reduced cost, higher customer satisfaction and loyalty, reduced compliance risk, higher sales and a host of other benefits.

For more information on Aspect’s educated approach to AI in the workforce, visit our Aspect Workforce AI page here.