Unlike manufacturing robots – the industrial machines we see on automated production lines and in manufacturing facilities – service robots require artificial intelligence to operate in the real world and successfully perform tasks on request.
Results from an Oxford University and Deloitte study stated that 35% of current jobs in the UK are at risk from automation – and that’s not just in the manufacturing industry. For careers related to customer service, there’s an even higher figure, with around 91% of jobs predicted to be lost to automation. In call centres, the realistic likelihood stands at a pessimistic 75%.
With such high figures, it’s no surprise that artificial intelligence, automation and robotics are often mistakenly thought of as the enemy of human workers. However, the vision of a robotic dystopia so frequently portrayed through science fiction is far from reality. In truth, many tasks that are managed by automation could be described as ‘back office’ jobs, rather than the tasks we associate with human employees.
For the most part, automation in the contact centre has been limited to managing menial tasks. Document management, data capture and cross-referencing have been taken out of the hands of contact agents and are now being managed by automated software. However, as technology advances, more complicated tasks may soon be within the capabilities of machines.
New developments in AI make the technology increasingly attainable for use in business applications. However, regardless of the name, AI does not necessarily mean the machine holds intelligence that is superior to a human worker. In fact, the way most business applications use AI is far from the sci-fi inspired vision many of us hold. There certainly won’t be self-aware software taking control of our contact centres any time soon.
Instead, most AI in business applications would be described as ‘soft AI’. This kind of AI can give the impression of intelligence by drawing information from datasets that are already available. Natural language processing, as an example of soft AI, is the term to describe the way computers can analyse language as it is spoken and written by humans – think how Amazon’s Alexa interacts and converses with humans seamlessly. Machine learning is another form of soft AI; this simply describes the way that computers can effectively ‘program’ themselves by adapting to changes in data.
Despite advances in natural language processing, AI is not good at reading sarcasm or human emotions. The technology lacks the human empathy and emotional skills to deal with complex customer service enquiries, such as complaints. Due to this flaw, the technology is incapable of replacing the human contact agent. Instead, automated technology offers contact centre agents the time to deal with more complicated customer service queries, handling the time-consuming administrative tasks that do not require human intervention.
SoftBank’s service robot Pepper can use facial recognition to pick up on sadness or hostility, and voice recognition to detect emotions — and it’s actually quite good at it.
However, the functionality of Pepper is limited to these very basic interactions and could not be used to manage a complex customer query. As it currently stands, artificial intelligence is used in automation to aid the human contact centre agent, or in the case of Pepper, as a fun novelty. Contact centres needn’t fear the rise of the robots just yet.
Stephen Parker is founder and chief executive of Parker Software