Artificial intelligence has changed customer service, but human touch is still important

Decades ago, the speech of telemarketers was stigmatized as being standardized and, often, full of gerunds. Since then, the traditional Customer Service (SAC) has changed significantly, with the use of new communication channels, such as WhatsApp , and technological resources, such as Artificial Intelligence (AI) and virtual customer service robots, or chatbots. However, even today, the reality of companies varies greatly and there are many challenges in delivering a satisfactory experience to consumers.

For Maurício Morgado, coordinator of the Center of Excellence in Retail at Fundação Getúlio Vargas (FGV) , high employee turnover is one of the issues that hinder the service provided by call centers. "It is difficult to maintain a good level of service because, almost always, the situations are complex and require training. So, one of the biggest challenges is keeping employees working, trained and informed about the company's problems and characteristics," he says.

The expert also highlights that the use of AI can be very useful, as long as such solutions work properly, which, according to Morgado, is not yet the case in many companies. "In their quest to reduce costs, they are seriously damaging the image of customer service. They have hidden telephone numbers, the famous 0800 numbers, and forced service through electronic systems. It is very difficult to reach a human [attendant]. And, when you do, you expect the service to be human too."

Journey

Getting to speak to real people was the difficulty that 40-year-old baker Virgínia Oda had with the customer service department of a car rental company. A few weeks after signing a long-term contract, her car was stolen. The car was found the same day, but was held for evaluation and maintenance. Then began a journey of countless attempts to contact the company, every week, for 40 days, to get the car back. "The primary channel would be WhatsApp. However, from the moment the theft occurred, our contract was suspended. So, I entered the CPF of the owner and the system didn't even give me the option to start the conversation," she recalls.

According to her, the only topic at the call center related to accidents was to report robberies or accidents, not to follow up on a problem of that kind. "So, every time, I had to type in a random option on the menu to get them to transfer me to someone. Furthermore, at no time did we have any active communication from them," says Virgínia.

According to Maurício Morgado, from FGV, regardless of the service channel, what matters is solving the customer's problem. "What is sorely lacking in customer service is empathy. Often, this does not happen because [the service] is bureaucratic. Systems and people have to have a certain flexibility to meet a consumer's needs," he observes.

Miscommunication and conflicting information at the call center caused a lot of stress for retiree Lidia Suzuyo Shigeno, 80. She contacted her telephone operator to exchange the metal cables for fiber optics and asked to transfer her line from an individual to a legal entity, at no extra cost and keeping the same number.

The operator said it would be possible, but she was charged a fine, lost her old number and had her phone on silent for weeks. Eight calls were registered with the SAC and several visits were scheduled. "Over the phone, they spoke as if they were going to solve the problem, but after I hung up, I waited and waited, and no one showed up. Or they sent technicians to my house who said they only worked with fiber optic cables, not metal ones," recalls Lidia.

Real intelligence

Positivo Tecnologia, a Brazilian hardware and IT services company, saw its customer service numbers improve significantly after implementing an Artificial Intelligence system that interacts with customers through menus and voice commands on its WhatsApp channel during the first half of 2025. According to the company, the percentage of customers who had their problems resolved without having to speak to a technician jumped from 28% to 43%. Average service time fell from 40 to 19 minutes, while average waiting time fell by 40%.

However, the introduction of generative AI was done gradually and in a planned manner. "We started studying this issue about three years ago, with 'bots', which were semi-automation. And we realized, right away, that if we didn't have [a mapping of] data and flows very well done, we wouldn't have reliable automation, it wouldn't work", says Marcos David Santos, Director of Operations and Services at Positivo.

"The attendants spent a year 'training' the Artificial Intelligence, 'teaching' it about our processes and products. We created a very rich and solid database and, only then, we started to put the AI ​​in front of the customer, talking to them, which is what happened about four months ago."

At the same time, the company is also investing in telephone support: eight years ago, the operators at its Customer Service Center were replaced by specialized technicians. "The entire team that serves customers has in-depth knowledge of the products and solutions. Therefore, they can resolve more than 98% of problems at the first level [of support]," says the director. "As a result, the resolution rate is high and we have low turnover, because these people have a better position and are better paid."