From the age of emails to chatbots

In 1971, computer programmer Ray Tomlinson sent the first email message to himself, probably not knowing that his discovery would become a key communication tool for generations to come. At the time of the millennium, if a business did not have an email address to communicate with its customers, it had not followed the trend leading to the era of digitization of communication in time.

Today, at a time when consumers are looking for instant service for their daily needs and transactions, the corresponding challenge is to support live conversation, known as chat. Chat comes to meet the need for short and personalized service in real time, from a specialized support agent. Analogous to the case of email, which we now all take for granted as a way of communication, so also for chat, it took a long time for businesses to understand its value as a communication channel with their customers. While chat first appeared as a form of communication for financial transactions (e.g. in e-shops) in the early 2000s, it took at least fifteen years to become a trend, initially for commercial transactions of consumer goods, but also in the service sector, with noteworthy personal experience in the field of tax advisory services. 

What comes next after chat?

In the age of rapid technological developments, the next step has already taken place with artificial intelligence. Chat applications are already being replaced by chatbot applications, which perform requested functions imitating human written and spoken speech as best as possible. 

A recent and very popular solution is a well-known chatbot that was developed based on both supervised and reinforcement learning techniques and quickly attracted the attention of the public for its detailed responses in a wide cognitive range, but also for the fact that as a machine learning model, it is called to manage large number of simultaneous users. Built on the same technology is a second, equally popular platform that uses natural language processing, machine learning and legal data analysis to automate and improve various aspects of related work. 

Of course, the effectiveness of such tools remains to be judged in practice, while the first signs of doubt are already being expressed. For example, in the case of the aforementioned chatbot, the comments are mainly about the outdated information it has been seen providing as it is fed information up to date by 2021, while reports say that a more up-to-date version of the app was recently made available to limited subscribers only. 

It is noteworthy that in a question raised in the application itself regarding the efficiency in the field of tax advisory services, the application replied that it does not have access to a database of tax regulations or laws but nevertheless, it can provide general information on tax matters, although these may not always be completely accurate. And finally, the application advised users for related questions to address specialized tax advisors! 

The fact is that the effectiveness of chatbots is probably still quite far from expectations, as human intelligence is always at a higher level. Of course, there are already reactions regarding the ethical dilemmas that arise with the development of artificial intelligence and the risks posed by the lack of a relevant regulatory framework. In fact, Italy, the first in Europe, recently banned the use of the application, at the request of the Italian personal data protection service, citing user security reasons. This decision has been preceded by China, Iran, North Korea and Russia. Despite the backlash, it is probably a foregone conclusion that proponents of human communication will have to accept technological advances. And especially if in the future it turns out that, given the continuous training of chatbots from their own conversations with users, their efficiency improves exponentially. 

Antonia Plataki is Senior Manager, Tax at PwC Greece and responsible for the Chat Tax service.