The replacement of people by machines in the attention to the consumer is a bad news for users.
In the current times it is increasingly difficult to have a phone conversation with a person of flesh and bone. Almost every time that you need to talk with the bank, with the doctor or any other service, the more likely it is that you welcome a helper automated apparently designed to avoid talk to someone who actually works for the company. This state of things may get worse in short due to the generalization of the chatbots.
The chatbots are programs of artificial intelligence often employed in applications or transmission services of messages. They are designed to answer people’s questions as in a conversation instead of just give you directions to find information, as do the search engines. Companies such as Uber, Lufthansa and Pizza Express are already used to respond to customer queries and record bookings, and many others are on the way to do it.
These virtual agents have the ability to improve some aspects of customer care, and, of course, are easier to use than the automated telephone systems that are difficult to understand even your basic personal information. However, they are also an obstacle that separates the users of a human being able to give real answers to difficult questions, and ‒fundamental‒ to show the compassion and the willingness to tend to be the basis of a customer service quality. It is possible that the chatbots be responsible for the clients and companies they will discover the basis of stumbling blocks.
For many companies, automating the customer service, or at least part of it, is a tempting idea. In this way not only manages to reduce the exposure of employees to many of the unpleasant situations own of this work, but it also helps to sift through numerous problems with current or no importance before attention is necessary, the more expensive a person’s. This could enable firms to reduce their costs, while it would serve to calm down the customers that only need simple solutions to common problems.
However, to replace employees on human artificial ones is not as simple. For a start, despite the advances truly amazing in the recognition and translation machines, the language, with all its variants and errors, remains a matter a huge can of worms. The automated agents are still too incompetent and little sensitive to him, and in the case of certain problems, it would be difficult or impossible to communicate with them.
Good, but not enough
The talent is the ability to achieve good results, and the domain, to solve a more difficult situation than normal. To handle exceptional situations is an art, and often the quality of a customer service have to do with cases that are unusual or unexpected in the involved clients, potentially angry. While the virtual agents can provide answers to basic questions in a convincing manner, the artificial intelligence is not yet sufficiently skilled to deal with cases of atypical and exceptional.
It is possible that, in the beginning, businesses do not see it as a problem, as the automation of the care introduces a way to separate customers whose service requires an additional effort. Just put in contact with an employee human to those people whose problems confuse the robot. However, it is likely that, as you go through the frustrating process of talking with a computer puzzled, the client is angry even more so with the service.In the long run, the result could be that this is seek another provider, in particular if you find it difficult to get a wizard a human to come to her aid when the robot is unable to do so.
I experienced a variation on the theme when you try to call a taxi after a train suffered a breakdown. I had the phone number of a company in the area and I called. They put Me with an automated service unable to recognize the pickup address is in any of the modes of naming and expression that I could occur.
By some happy chance, they put me with a human agent, but, before he had time to explain the predicament that I was in, this told me that I was wrong with the reservation system, and the loop of hell started again. This sad story ended with a hike, a lucky pick for a black taxi driven by a person in an area is otherwise absolutely deserted, and the oath of that, going forward, would prevent the company first whenever possible.
The automated services can take care of the cases, but are still unable to adapt to the exceptional circumstances or, at least, recognize when it is necessary the flexibility of human intervention. From the point of view of the client, the problem goes even further. Some situations do not only require the human capacity to understand and to solve the problems, but a dose of compassion and empathy.
It is possible to program a virtual agent to adopt a particular style of interaction, but, in unexpected situations or difficult, will continue resulting in strangely inconsistent. Today, the research of the artificial intelligence does not have a sheet path functional that allows you to apply some semblance of human compassion in a convincing way.
Sometimes the angry patrons in need of a kind word and the chance to express themselves with someone willing to listen, as well as ‒or sometimes instead of‒ that will solve the problem. And, often, the quality of customer service depends on the goodwill gestures made at your discretion for a specific employee by following their own feelings of empathy, rather than a series of fixed rules.
This is something very difficult to reproduce by artificial intelligence, because it depends largely on the context of the situation. In my opinion, the understanding of the context remains one of the most elusive and pending resolution of the discipline, and is likely to remain so for a few years.
Despite this, it seems that the promise of cost reduction, in addition to other benefits of automation, they are so attractive that, in the next few years, chatbots, and other services of artificial intelligence aimed at the customers are going to continue to expand, without hesitation. What is more likely is that, in the medium term, the result is a treatment of the complaints even more technocratic and less flexible. Or even worse. As the algorithms are refined, the process of decision making can become opaque and leave very little margin to the intervention of a human supervisor.
If we do not want this to happen, we need to be aware of that the way the assistance is not paved with good intentions, but that is based on the understanding of how limited that is by today’s artificial intelligence to understand the contexts, the exceptions and the human condition.