…When machines or human share responsibilities
I want to believe that by now we have all accepted the fact that we are all going to have to work side by side with robots—if not now, then very soon. In whatever form they come, virtual or physical, robots are going to be our co-workers. And as our robotic colleagues become much more intelligent, thanks to artificial intelligence (AI), we have no choice but to accept their importance in the growth and increased productivity of our businesses.
I believe it is easier to embrace a machine as a valuable colleague when one sees the quantum of work the machine does. In many sectors, AI and humans are already working side by side. Whether aiding medical staff to do a better diagnosis, helping shoppers have a better shopping experience or even detecting and informing farmers as to where weeds are more likely to grow. It is safe to say that these smart machines are here to stay.
As we work with these machines, there must definitely be a need to decide who will do what. Humans and robots both have their strengths and weaknesses. Ask those who are currently employing robots and other artificial support, and they will enumerate a litany of advantages these robots have over human workers. Robots do not get sick and thus they will not take sick days off. Robots will not have a problem with a colleague and consequently create unnecessary tension and drama in the office. In fact, these employers will go as far as considering their AI and robot employees as “perfect employees”.
In a recent interview on the BBC World Service, a restaurateur who had a robot helper and was even deciding to go for another one, said the best thing about the robot was that she had just rented the robot and so she pays the original owners a monthly fee. This, she said, made the use of robots as workers so cost effective. The labour cost implications and the efficiency of these machines make their rapid adoption almost a no-brainer.
All these makes it clear that businesses will continue to adopt AI and robots more and more. This is not to say that robots will do away with human workers totally. True. There are many examples of AI agents replacing human agents at the front line. But not to worry. You and I will still have very important roles to play in the workplace. At least, that is what the experts tell us.
What all this means is that businesses must necessarily find a way of dividing the jobs in line with the strength of the human worker and their machine alternatives. For instance, the more monotonous and mundane jobs that demand specific quality standards can be given to the robots while jobs that demand creative use of the mental faculties can be assigned to the human workers. Jobs that would require advanced processing capabilities are definitely the domain of these AI agents while those jobs that demand an emotional input can be set aside for those with flesh and blood.
As we divide the various roles among human and robotic workers, it is important not to lose sight of the fact that one of the very important job roles is that of interfacing with customers. It is one thing to have machines working behind the scenes and getting things done effectively and efficiently. It is however a different ballgame when these machines have to deal directly with customers. It is expected that those who interact with customers must ensure that the experience of these customers are so pleasurable that customers keep coming back for more. It is a fact that robots as well as humans can interact with human customers. But the question is, which of the two, humans or machines, can do the job of dealing with the customer better? Which of the two do customers really want to be dealing with directly?
Thankfully, there are some attempts made at finding out the answers to these questions. A recent study threw some interesting light on the how businesses can deploy robots and humans, especially when it comes to communicating with customers. This study was published in the December 2021 edition of the Journal of Marketing. The study was emphatic that the way customers respond to information or messages from organisations is dependent on whether the customer interacted with a human being or a machine.
It is a truth that an organisation will regularly have to send information to customers in the process of offering service to these customers. It is also true that these messages might not always be what customers would be expecting. For instance, I am not aware of any customers who are delighted with unexpected increase in prices or delays in provision of services that have already been paid. Other such messages might include information regarding service failures, product defects and subsequent recalls, product stocks running out, among others. These are messages that customers generally resent.
According to the Journal of Marketing study referred to above, when a firm has to deliver any of these negative messages to a customer which is less than what the customer expects, customers respond better when it is an AI that delivers the bad news. Customers will let an AI agent get away with giving a not-so-pleasant message—something that these same customers will not accept from a human agent.
It is also important to understand that it is all not bad news when it comes to dealing with customers. On many other occasions, the news can be very good. For instance, upgrading a customer’s product or service as well as refunding monies spent on products that failed to deliver can be considered good news. Replacement of faulty products and giving customers surprise gift items also qualify as good news. In such situations, according to the above-mentioned study, it is smart to get a human being to deliver the good news, rather than an AI agent.
The explanation given by the authors of the aforementioned study was that people did not attribute ulterior motives to AI. Customers believe that AI agents are just machines, with no selfish ambitions. Therefore, when the AI makes an offer, it is not trying to cheat the customer, unlike a human being that might have the capacity to cheat, steal and lie. In short, the study found that customers are more forgiving of non-human agents when things do not go according to plan. In the same vein, when an AI agent delivers good news to a customer, customers are generally less appreciative of the AI. The assumption is that the AI does not deserve any special recognition since the decision is not as a result of an independent decision by the AI.
I also want to believe that another reason why customers might let the AI agent get away with the not-too-good offer is that customers might feel they might look silly trying to argue with a non-human. Customers might feel there is really no use arguing with a machine into which everything has been pre-programmed.
The title of the aforementioned study is “Bad News? Send an AI. Good News? Send a Human”. The takeaway from the study is that businesses can take advantage of this phenomenon by attributing human characteristics or behaviour to their AI agents. If the news is good, the humanness of the AI can be emphasised so that the AI agent can be made to deliver the news. If the news is not too pleasant, then the anthropomorphization can be reduced so that the customer is aware it is just a machine.
This need to make AI agents more human can explain why a number of AI service agents all over the world are given real human names. In Ghana, there is Kukua from Fidelity Bank Ghana. Then there is Abby from ABSA Ghana. Vodafone Ghana has TOBi. CFAO Equipment Ghana Ltd also introduced a virtual assistant chatbot, named Charles. Araba is the name given to the AI assistant from Hollard Insurance Ghana. These organisations are using these names to attribute much-needed human characteristics to their non-human agents.
As a matter of fact, the above-referred study stated that the attitude of customers towards intelligent machines change when these machines take on more human-like features. Robots that tend to look more like human beings, with human body structures and having facial hair, are said to elicit favourable responses from customers when the offer is better than expected.
In all of this, it is important for organisations to do some deep thinking before jumping on the intelligent machine bandwagon. The specific situation and context of the organisation as well as the attitude of its customers to robots, AI and smart machines must all be taken into serious consideration before any such venture.
The truth is that it will take some time before we truly come to terms with the real challenges of placing human and AI agents side by side at the front line. Time will unfold some other challenges that we might not have anticipated. We can however be thankful that, in the very least, we now know which of the two to send when it comes to delivering of good or not-so-good messages to customers. Don’t forget, when it is bad news, then it is good to send the robot.