The Robots Are Coming: Look to Your Nearest Neighbour – Jeremy Hamill-Keays, Product Marketing Manager at Teleopti, considers what the future of digitalization and AI means for WFM. What could you learn from your neighbours?
There is a lot of news and noise around the subject of Artificial Intelligence (AI). Some industry experts are predicting the end of our culture as we know it, with humans spending their ever-increasing lives freed from work and instead pursuing a life of leisure and perhaps, poetry.
I don’t have any idea if that will happen, but I can say with certainty that things will be different. AI is a broad definition. Some may consider landing a plane in heavy fog as a pretty intelligent system, and one that can easily perform better than human counterparts. Computers were beating Chess Masters decades ago and stock markets now have rules in place to stop computers causing high speed mayhem due to lightning-fast trading. But is this really AI?
Everyone has their own definition of AI. Mine is “a system that can analyze data and provide a result based on history and which can learn to produce an optimal result”. Wow, that is a lot of words.
What do I mean exactly?
Intelligence is based on the ability to learn. For example, a computer is tasked with sorting 100 pictures into dogs or cats. First it is shown pictures of 10,000 cats and dogs, and tries to select the correct category. The test data is already categorized so the service can be trained, and it may run through the test data multiple times, slowly enhancing its algorithms. Once complete, the system can now be shown new pictures and make a pretty accurate guess based on its training. This same technology is already used to spot cancer cells, with a current accuracy rate of 86%, it is more often right than a trained human. This all came from scientists collecting tens of thousands of high resolution images of pre-cancerous and cancerous cells in order to ignite the machine learning process; the AI algorithm was then able to discern cancers from highly magnified pictures within just a second. Essentially AI needs a mass of data to form its memory and shape its learning, indeed a bit like a human – we learn our behaviour and make decisions based on experience.
So how could this affect WFM?
AI comprises of many techniques, and offers a whole host of opportunities. I believe one of the most interesting ones for WFM in the contact centre is having the AI option to match a new traffic pattern to the nearest matching patterns in the centre’s stored history. If we configure the system to use the 10 historical morning traffic patterns best matching today’s morning pattern, the system will find the best 10 and then use them to predict what the afternoon will look like. As more historical data is added to the system, the output is also updated, the system is adapting to new information and learning the patterns that fit best. This technique is called “Nearest Neighbour” and comes from research into image recognition. It blends statistical methods with “Artificial Intelligence” and is a technique that Teleopti is considering for our Intraday Automation functionality.
Neural Networks have been available for a long time, but with the production of cheap computer chips designed to handle complex algorithms for digital gaming, developments have really accelerated. These networks have the capability of handling many different factors and complex situations, using very large data sets with very large neural networks, the results are astonishing. This is where much of the news is, so called Deep Learning. A good example is being able to formulate speech for a large number of words, this is difficult, even for humans. AI systems today can cut up and reform word samples in order to pronounce new words, and this ability will make a difference to how calls are handled in the contact centre with customers being able to call an intelligent robot and talk. AI systems can watch humans at work and adapt their own algorithms to learn. They can also help humans learn, pointing out to agents when the customer gives a buy signal, or if a supervisor should come on to help. Today the systems are large, but you only have to use an iPhone with Siri, or Amazon’s Alexa, to understand where the future is headed.
The technology is here and will only grow bigger. What does it mean for employees in the contact centre? Well, in the foreseeable future, AI will be able to handle fairly simple tasks, but more complicated processes will require competent humans. Having spent time, effort and money training an agent, it is important that the agent is engaged and also loyal. WFM systems will need to balance cost efficiency and optimization with human factors such as flexibility of shifts and visibility to deliver agent engagement. But will we replace agents with AI? In some instances, yes, and this should lower costs, allowing for more affordable goods which result in more customers and more calls, some of which are complex and will thus require more agents… A cycle it seems. A cycle of prosperity. Rather than fearing the robots coming or seeing them as a way to cut back on agent staff, it can instead be seen as a chance to challenge practices, support customer service staff and processes, and in general, grow.
Jeremy Hamill-Keays is Product Marketing Manager at Teleopti