Table of Contents 64q43
On this May 26, 2021, the IBM invited experts to talk about artificial intelligence, its development and use in creating virtual assistants based on the IBM Watson, the cognitive services platform for business. With applications that range from helping to combat fraud and cybersecurity to compliance and service to employees and customers, this technology has been increasingly present in everyday life and has evolved exponentially. 35s2f
coordinated by Juliana Cayres September, head of communication at IBM, the videoconference was attended by specialists Claudio Santos Pinhanez, conversational intelligence manager at the IBM Brazil research center, and Marcela Vairo Pasqualetti, Director of Data, AI Apps and Automation at IBM Brazil.
Virtual assistants, their demand and current situation 53702b
To open the conversation, we were told that the adoption of AI for virtual assistants based on the platform IBM Watson Assistant doubled between January and December 2020, in an initiative to meet the need to accelerate the process of digital transformation of companies. Partly a reaction to market changes generated by the new coronavirus pandemic, the platform has been helping organizations from different segments to manage any spikes in consumer demand and improve customer service and satisfaction.
Taking the lead, Marcela Vairo commented on the evolution of virtual assistants, which are becoming increasingly friendly. This evolution brings advantages in the relationship with the end customer, with the example of the use of AI in cars: when identifying a problem in the car, the assistant not only informs the consumer but also the manufacturer, helping to solve it quickly and also helping to marks to understand, with statistical data, the frequency of problems and the conditions in which they occurred.
“Customer service is the biggest driver for the use of AI in Brazil.”
Marcela Vairo Pasqualetti, Director of Data, AI Apps and Automation at IBM Brazil
Speaking of which, according to a survey – the Global AI Adoption Index 2021 – Brazil has been at the forefront of AI adoption among Latin American countries: 40% of Brazilian companies have already adopted it AI solutions, against 21% in Latin America. An example is the Original Bank, where 93% of the service is done entirely by artificial intelligence. Another data revealed by the survey informs that 60% of IT professionals in Brazil reported that business needs are driving the adoption of AI in companies, and 37% of respondents said that due to the pandemic, the focus on customer service has been intensified.
In of customer service, the IA It also helps a lot in the fight against fraud. By quickly identifying hacking patterns or other types of intrusion, it helps speed up service by reducing the steps needed to meet requests, improving cybersecurity and compliance for banks and various companies. These AI business needs, which extend consumers' digital experience and ensure business continuity, are necessary to be able to serve s very well even during the most critical periods of social isolation.
The internal use of technologies AI and Virtual Assistants it is also present in companies — dealing with employee profiles, modifying and managing data with much greater ease compared to old systems, much more complex in of use and loaded with information and screens.
“Business and working relationships have changed. Suddenly, everything had to adapt to online and this whole migration from real to virtual will increasingly need the of AI and the maturation of technologies that were already expanding, such as data analysis and automation.”
Marcela Vairo Pasqualetti, Director of Data, AI Apps and Automation at IBM Brazil
Past, future and challenges of AI 1k1y18
Claudio Pihanez, a veteran in the study of artificial intelligence, gave a good overview of how technology has evolved since its inception and what challenges it has gone through — and will go through. He related the evolution of neural networks in the IBM Watson turn, in 2011, when the system won the Jeopardy, an internationally famous quiz game where you only have to answer with other questions — and the Watson he did it masterfully, using human language.
Since that time, the IA struggle with what is called Similarity Problem: We humans are good at spotting patterns, and we apply that to our technology, but we don't know how this system works in our brain. Programming this into an AI has been a challenge from the beginning precisely because of the complexity of this pattern identification. The evolution of this, recalls Claudio, could be observed in the analysis of the tweets made in relation to the Soccer World Cup 2014, where millions of messages were analyzed by Watson to understand what feelings were being expressed by the population, what they were, and at what intensity. This was one of the first major – I mean, mass – real-time uses of the technology.
Another challenge still present is the databases in Portuguese. While abroad the English language has a lot of information to work with and huge databases, in Brazil it is necessary to feed the system not only with the standard language, but also with its variations - an example used is the word “steppe”, an element of automobiles. that have different names in different places in Brazil, and that need to be ed for so that a virtual assistant understands and can communicate with the client in an eventual problem involving the object.
The adequacy of data and its integration with systems is also important in this regard. There is what is called AutoAI, the automation of this machine learning where artificial intelligence itself manages to generate and feed its models, but it is necessary to start from somewhere. Assembling teams to form these systems, working on processes and data is necessary so that automation of this level can be achieved. Natural language processing is still under-resourced and under-staffed, unfortunately, according to Claudio.
“Artificial intelligence is not about changing people, it is about doing different things”
Claudio Santos Pinhanez, Conversational Intelligence Manager, IBM Brazil Research Center
One of the concerns that exist about artificial intelligence is that it could end up replacing workers in the services provided, and Claudio comments that it is not a replacement that will occur, but a change. While many services are automated, professionals are required to understand and handle cloud (“cloud”, in free translation), to make a contribution between AI and business models, among others. There is a lack of trained professionals for this, according to experts.
Marcela Vairo even comments that AI is not used to do more of the same, but to do it better: to enhance services. In addition to creating smarter companies, it was stressed that the whole process also helps with transparency. With the Brazilian General Data Protection Law (LGPD) (General Data Protection Act), this has become even more necessary.
Claudio Pinhanez concludes by citing the following challenges in improving the IA and virtual assistants based on IBM Watson. In addition to similarity-based processing – that is, pattern recognition – the next step would be the inclusion of a “reasoning” of its own, a network capable of generating its own novel conclusions. To this end, concepts such as neurosymbolic computing mentioned, which are complex and still have a long way to go. Furthermore, there is a lack of appropriate hardware for neural networks, such as more analog technologies, such as optical processing. In view of the future is even quantum computing, whose research is currently very promising.
Use of AI and virtual assistants in Brazil 5e146p
“As organizations move into a post-pandemic world, IBM sees AI as a great ally to help companies automate processes, increase trust and understand the language of business. We believe these capabilities are critical to transforming the customer experience and helping businesses emerge stronger from the pandemic.”
Claudio Santos Pinhanez, conversational intelligence manager at the IBM Brazil research center
As we have already said, Brazil is a great pioneer in the use of neural networks and AI. One of the examples is the Bia, the virtual assistant of the Bradesco, which was cited by experts as one of the most advanced ever seen. already the Inter, a digital platform for financial and non-financial services that aims to simplify the lives of Brazilians, created in 2018 the Babi, a dynamic virtual assistant based on IBM Watson Assistant, capable of resolving different types of requests in real time. With the wizard based on Watson, a bank, which has 11 million customers, has already achieved a reduction of 16,5% in the cost of service, 45% in real customer retention and total savings of R$ 16 million in just one year.
“The use of AI has already proved to be a great ally for us to bring an increasingly better digital experience to our customers. With the pandemic, technology has proved to be essential for us to keep our communication channels active and a complete and efficient digital journey. Our intention now is to integrate Babi with WhatsApp messages, making digital interaction even more agile and accurate.”
Priscila Sales, CMO of Inter
A TIM Brasil, on the other hand, worked with the IBM to streamline customer service workflows. Through the partnership, an AI-powered virtual assistant was deployed to help handle the high volume of calls in an effort to maintain customer satisfaction, reduce costs, and free up call center agents for higher-value work.
The company's virtual assistant uses the IBM Watson Assistant em IBM Cloud to process natural language interactions, responding by voice in real time to customer questions or concerns about issues such as plan benefits and bill payments. Four months after increasing human customer service agents with the virtual agent, TIM's contention rate increased to 75% and first call resolution increased by 85%. For those who were tired of having to spend hours on the phone pressing buttons to get the call disconnected before solving the problem, this is a blessing – and a huge benefit for both sides.
Asked if we can know where we are in the evolution of IA and virtual assistants, Claudio Pinhanez gave an exciting answer: this is just the beginning. It is still possible to identify an artificial intelligence behind a conversation, for example, because human interaction is very complex and difficult to reproduce; however, calculates Claudio, in a matter of five years or less, we will be at a point where the conversation will be natural, and for this to be possible, the evolution of technology will have to increase a lot – and for the better. We still have a lot to evolve and a lot of applications to do with what AI can provide us.
So, excited about the future of technology? Tell us in the comments! And to keep getting inside the AI and technology news, keep an eye out here at showmetech.
Source: IBM
