For many, artificial intelligence exists only in films or conspiracy theories about the end of the world. Although it is already present in some areas of the labor market, such as marketing, for example, it is still considered by many as something futuristic. However, in reality, the use of AI is increasingly affecting the way customers and companies communicate.
The challenges and impacts of this new way of communicating were some of the topics covered in episode #261 of Podcast-se, the podcast of Comunique-se, which addresses topics related to corporate communication, press relations and digital marketing.
The conversation was mediated by Cassio Politi, specialist in content marketing and founder of Tracto. As guests, Marcel Rosa, head of design and partner at Alana AI, and André Calvente, marketing manager also at Alana AI.
Check out the chat in full.
The impact of AI on the brand x consumer relationship
With the emergence of this need to include AI in communication, right at the beginning of the conversation Cassio questions how artificial intelligence is going to change or is changing the relationship between brand and consumer.
To make the explanation clearer, Marcel gives a background on the neuroscience perception and action cycle, the basis of AI in marketing. Through the collection - reasoning - action cycle, artificial intelligence is able to transform all the data collected into insights, which are used so that an action can be taken - in the case of marketing, so that the right message reaches the right person at the right moment.
Three steps of the learning cycle
Collection Reasoning Action
Rosa explains that, when done correctly, this process is used as a problem-solving tool. This leads to an approximation between brand x customer. All through direct and personalized communication, without making the process repetitive or tiring.
Reading recommendation: Artificial Intelligence for e-commerce: applications and benefits.
What is AI's biggest marketing challenge?
In 1996, the first clash between human and machine took place. Garry Kasparov, one of the greatest chess players of all time and Deep Blue, an IBM supercomputer programmed to play head to head with any chess player in the world. Although the first match was won by Kasparov, Deep Blue took the rematch.
The game generated a lot of controversy, and Kasparov even claimed that it had been a cheat, for Deep Blue had made decisions that would be considered "incoherent" for a machine. This even generated a documentary, called The Man vs. The Machine, created by ESPN.
With this example, Politi observes that, in the world of chess for example, although there are multiple moves, the variables are finite - and questions whether the great challenge of AI wouldn't be precisely to have to deal with infinite possibilities and scenarios that multiply all the time.
“When we analyze the interactions that happen, whether on social media, or anywhere else, we start to see patterns.” Rosa says. Thus, there are several challenges that artificial intelligence still needs to overcome, but this is not one of them.
One of the problems that companies face is the failed attempt to humanize a service by hiring humans, but always wanting to keep the same script. This attempt, ironically, ends up making the process robotic.
With Alana AI, for example, this does not happen. From the moment that the subject and the feeling are identified, it is capable of generating multiple responses, interactions, and micro changes in the texts, not behaving in a robotic manner and consequently excluding any difference between the service provided by the human and that provided by AI.
During a play in an interview between Patricia Meirelles, Patricia Meirelles, ambassador of Alana AI, and Marcio Kumruian, founder and CEO of Netshoes, Marcio cannot differentiate which answers were given by a human attendant, and which were given by Alana AI, for example.
Rosa explains that, after hiring, during the onboarding configuration of Alana AI, it is possible to make the answers more fun, more objective, configure the AI to make sales, increase engagement or even answer questions.
AI X Humans in customer service
With that in mind, Cassio questions to what extent the use of humans in customer service is unnecessary. Marcel says that AI is a tool, and even with all its updates, it still works only to a certain extent.
André also points out that Alana AI is a co-worker. And from the moment that it realizes that it cannot solve a problem, the interaction is passed on to a human agent, precisely because humans are capable of having feelings, such as empathy, that are often necessary to deal with those on the other side, which a machine cannot achieve.
Politi then points out the existence of the two types of AI, one being general and the other narrow. This subject was also one of the themes of the first episode of the Inside Alana Podcast, where Marcellus Amadeus, CTO of Alana AI, explains more in depth about these differences.
In summary, AI General would be the technology considered ideal, one that resembles what we see in films and series about artificial intelligence. I would be able to reflect, make decisions and act according to context, feeling and environment. Already the Narrow AI is the format of artificial intelligence that we have today, in which they are able to perform specific tasks and assist humans in relatively simple tasks.
Reading recommendation: How artificial intelligence is innovating customer service chatbots.
Will robots have the ability to replace copywriters?
In September of this year, The Guardian surprised its readers by making an article written entirely by a robot. This raised debates about the future of journalists and content creators, fearing that they could be replaced.
When asked about it, Marcel and André agreed that replacing 100% wouldn’t be possible. This type of task involves many processes that, although they can be simulated by an AI, an AI doesn’t go very far. They are just a simulation, and that would end up being reflected in the text.
"The role of artificial intelligence is to interpret contexts and information, simulate behavior and generate a result", says André.
With this in mind, it would be difficult for a robot to write a political text, for example, or to assemble articles without a possible prejudiced bias, since it works according to past information, which may belong mostly to a single group.
A clear example of this would be a person being approached by a police officer for driving above the permitted speed. If the only difference between these two people who committed the offense is color, the data provided for AI will most likely determine that blacks provide more risk to society, precisely because the number of blacks in the prison system is higher.
André mentions these problems that may arise if artificial intelligence assumes this responsibility, and adds the Blackbox phenomenon, which is when the AI comes to a conclusion and it is impossible to explain how, being untraceable, the place from where it took a certain opinion.
The topic was even debated at the International Symposium on Online Journalism, and the Associated Press put together a report on the subject. It is less about replacing, and more about assisting.
Simple tasks that normally require a lot of time, would be performed by these robots, automating everyone's time and work, not just journalists. By releasing humans from “simple” jobs, we will be able to focus on possible problems and situations that an AI would not be able to solve.
To know and understand more about these subjects, check out the Inside Alana Podcast, which already has two seasons. The first with the objective of introducing the world of artificial intelligence and its bases, with Regina Bittar, Siri's first voice in Brazil, as a presenter. In the second season, check out mini episodes that address how artificial intelligence and humans are working together for the benefit of society in several areas.