The platform, as mentioned, mimics human senses, in an integrated way, which gives rise to different services. IBM brought two practical examples to the presentation, considering that M.A.R.I.A. it is currently integrated into the system, in the learning phase and in the friends and friendly circle, and is currently learning the different Portuguese accents. Internal help aims to free workers for tasks that are really necessary, giving the robot repetitive processes. In this case the human makes the final decisions that were made by the machine.
For example, pledges and other risky processes are the platform's priority, seeking to understand what is on the screen and apply what you have learned to execute decisions, in terms of the internal use of the platform, that is, without interaction with the customer. Montepio points out that the system is in compliance with the directives of the European Commission and that all decisions and actions are recorded and can be audited when necessary.
The project was developed in Portugal, by the IBM teams in their nearshore of Tomar and Viseu, in an execution in Portuguese. IBM assumes itself as an outsourcing leader, since three years ago it has been offering native cloud to the specific needs of each client.
In external contact with customers, in Contact Centers, the robot already executes 72 customer intentions. The human, as a rule, does not like rules in conversation, because he wants to solve his problems quickly. The lack of patience leads to a failure of 50%, in general, of the success in the communication of the call centers. The activation of customer cards and account confirmations are tasks that can already be performed by the bot, and will serve as a pilot in its launch. Security issues were addressed and the company says that the bot has greater resilience.
From a technical point of view, there are components outside the bank, such as the attachment platform, for example, making the automation complex. In this case, the bot gathers the task list of the entire process, with cross checking in the middle. Considering that an improper attachment harms the bank's statute, this functionality has been rigorously refined. Montepio also presented a road map of automation, predicting that in less than a year about 80% of the work flow of its Call Center will be serviced by the machine. Analyzes made automatically also speed up credit granting situations. The platform has a challenge to scale and receive high flows of calls, as for example in campaigns, and above all to serve customers outside the conventional schedule.
For now, working internally, Banco Montepio expects to launch M.A.R.I.A. to the second week of April. At first, only by voice channel, through the Contact Center, but in the future via chat directly on the bank's website. Banco Montepio says that for now the machine will only work inbound, that is, from the client to the bank, to resolve doubts, but it is prepared, and eventually in the future it can become outbound, with the sale of products or suggestions to its customers. customers, namely advice on housing credit services, for example.
