-Hello! I want to know how much I have to pay - asks the user, in a hurry, avoiding one of the signs of admiration, the capital letter and the corresponding tilde.
-If you want to check the amount of your debt, enter your number from here: -responds the virtual agent, in a recognizable River Plate Spanish.
-ah ok, thanks and where can I pay? -continues the client, stripped of any concern for the spelling.
-We tell you the payment places enabled to pay your monthly bill according to the payment method you choose. Keep in mind that the time of imputation of the payment varies according to the means of payment used -contacts the chatbot, slipping some redundancy, it is worth mentioning it.
-Thank you!
-You are welcome.
Hundreds of thousands of times per day similar dialogues are played. In fact, a recent study by Endeavor and Everis shows that the income generated in Latin America by the direct and indirect application of artificial intelligence -which allows virtual agents to interact with people through electronic means- will rise from US $ 3. .200 million current au $ s89,800 million for 2025, with an annual growth rate composed of 52%.
In addition, the report states that 30% of the enterprises that use artificial intelligence in the region are focused precisely on the development of chatbots.
Founded in the Cordovan capital in 2012 and with offices in Brazil, Colombia, Mexico, a brand new in Spain and soon in the United States, Aivo is one of the leaders in that segment. In fact, it has more than 100 clients in 32 countries and serves more than 2 million people per month.
To make matters worse, has integrations with global players such as Salesforce, Zendesk and LivePerson; raised capital for US $ 1,900,000; and its sales of omnichannel automatic service solutions to the customer 24 hours a day, 7 days a week - using artificial intelligence - have already exceeded $ 15,000,000 along its trajectory.
"Our system consists of a set of artificial intelligences specialized in customer service, we create a conversational engine that has different layers, in the first one, spelling errors are fixed, regionalisms, emojis and voice messages are interpreted. to see with the natural conversations ", indicates Martín Frascaroli, founder of Aivo.
In Creole, on the one hand, the bot understands both the phrase 'I want to pay my bill' and 'Che, where I garpo the ballot'. On the other, the system has a memory stadium that allows you to understand what product it is, if someone asks about the stock of the last smartphone and later asks 'How much does it cost?', Without mentioning it again.
"This process provides the possibility of dispensing with technical staff and linguists for its implementation. Each bot is aimed at solving the problem and is not a flow of options," continues the entrepreneur.
As it could not be otherwise, the companies that implement this type of virtual agents seek to automate the attention of their customers, as much as possible increase their satisfaction and, above all, save costs. Everything, through a technology able to understand the intentions of the users and to solve their needs in few interactions.
In action
With six years of experience in the use of chatbots, the main virtual agent of Movistar Argentina is called Sofía and responds to the queries of the clients that enter over the Internet, from any device, about the home and mobile businesses of the company.
"Our products and services are very diverse and have very complex information flows.The biggest difficulty is the understanding about the consultations, since they can be written with or without errors of spelling, spaces and idioms.This is all part of the training and the capacity of the robot, "says Verónica Puentes, head of Digital Relations of the company.
In this regard, he points out that such implementation not only allows the company to increase the attention capacity of customers who do not want to call by telephone or need the assistance of a person, but also their satisfaction, in addition to having constant and immediate feedback to Feed back to the chatbot and other channels.
"Our virtual assistant solves 87% of the queries that come in. Only 9% is transferred to a person for their solution, the indicator of satisfaction of this channel is 84% ââpositive and 7% negative, while in the remaining 9% is neutral, and of the inquiries that enter, only 4% are then transferred to other means of care ", the executive completes.
In this sense, the founder of Aivo explains that the differences between one chatbot and another depend on the vertical market application, while the technology is the same.
"For example, in the case of telecommunications companies, most of the queries have to do with changes in equipment, plans or balances, and in e-commerce, the questions are about stocks, prices and returns. banking, about the balance of the credit card or the credits The interesting thing is that, the more users consult and leave their feedback, the more the bots learn, "says the entrepreneur.
Implemented in May 2016, for its part, Lucía is the virtual agent that responds to the requirements of BBVA Francés customers.
"It has links to certain contents and the contact form and includes tutorial videos for example about the bank's benefit program.For the client, the solution allows us to provide solutions 24/7, with empathy and natural language. great advantages is that it offers the possibility of deriving the query to different areas of the bank depending on the requirement ", illustrates Christian Petruzzi, its manager of the Customer Service Center.
In particular, Lucia solves an average of 200,000 queries per month, concentrated in 100,000 conversations, according to the executive. "From the beginning to the present, the average of positive feedback is 2.47% against a negative one of 3.93%, surpassing almost the entire industry, whose average is 1% of positive feedback against 5% negative. Some of these results are based on the constant updating of the content, based on the types of consultations carried out and the average time it takes to resolve them, "he concludes.
It is clear that the application of artificial intelligence developments is still very far from the replicators of Blade Runner. But, at least for now, it serves so that the complaints or claims of many users are not lost in time like tears in the rain.
