Artificial intelligence reaches the call center: how does the banking voicebot respond and what will it be able to do?

The BCP, since March of this year, has implemented a voicebot in its telephone contact center. This is a project that is pioneering in the sector and has the technological support of IBM. What does the tool consist of?

It works as a kind of virtual voice assistant or bot which uses artificial intelligence to interpret and respond to bank user queries quickly and efficiently.

More agile and precise attention

When calling, the customer is greeted by the tool that acts as a first point of contact. Instead of requiring the user to navigate traditional menus, they can simply speak out loud what they need. AI is designed to recognize and understand these requests allowing the system to process them without having to press numerical options.

The voicebot uses an advanced AI system that understands and classifies customer intent. If the query is informational (such as an account balance or credit card details), the bot you can answer directly. For more complex requests, such as those that the bot cannot resolve, The system performs intelligent routing that directs the call to the corresponding service area.

Thanks to this intelligent routing, if the voicebot does not have the necessary information or detects that the client prefers to speak with an advisor, it transfers the call to the appropriate group, saving time and streamlining the service. In addition, the voicebot has a system of Customer prioritization that detects signs of frustration or urgency and, in those cases, it passes the client directly to a specialized team for immediate attention. Indeed, the AI ​​detects signs of discomfort by recognizing certain types of words or the tonality of the interlocutor’s voice that alert it to the need for more personalized attention.

But what is the difference between this new system and the one previously used in the call center?

“Previously, the service system used an IVR (Interactive Voice Response), which guided the customer through audio options such as ‘dial one’ or ‘dial five’ to reach the desired information”Michael Campbell, leader of the BCP Contact and Sales Center, explains to El Comercio. What generated this? Normally, there would be many more passes to the advisor because the client did not always arrive at the correct option or did not know which option to check and chose to go directly to the advisor.

“Then there was the problem of customer abandonment. Many became frustrated at being stuck in the tree of options and around 50% hung up before reaching an answer. Today, with the voicebot, that dropout rate has dropped to less than 10%”details Campbell.

Generative AI: the next step towards transactional banking

Perhaps all this does not sound very revolutionary, even more so if we take into account all the “skills” that generative artificial intelligence has recently demonstrated, from the ability to maintain advanced conversations, to creating audios and videos practically from scratch and with impressive realism. However, for the bank, these advances are a sign that they are moving in the right direction. As Campbell points out, “With the previous IVR system, the average wait time for a customer to hear all the options and be directed to the appropriate area was about two minutes. Today, with the voicebot, that time has been reduced to less than 15 seconds”.

On the other hand, Álvaro Santa María, general manager of IBM, tells this newspaper that “the first phase of this project has been implemented using new generation traditional artificial intelligence, which allows the system to understand and respond to queries quickly and accurately. ”. However, according to the roadmap established together with the BCP, the second phase will include generative artificial intelligence. “This addition seeks to further improve customer service, allowing more personalized responses and a higher level of interaction to satisfy user needs,” he says..

MIRA: From ChatGPT to interactive apps: how artificial intelligence is changing the way we learn English

With generative AI it will be possible to move to transactional stage. This second phase will focus on allowing customers perform financial operations directly with the voicebotsuch as money transfers, payments and recharges. The inclusion of these functionalities responds to the objective of streamlining processes, improving the user experience and reducing dependence on human advisors in routine queries.

This advancement requires a high level of accuracy and security in each transaction, as more complex queries have to be processed and responses personalized in real time.

“It is not just about providing information, but moving towards a platform that can carry out transactions in response to the needs of each client, quickly and accurately,” expresses Santa María.

The fear of being replaced

The fact that AI has more and more autonomy to carry out jobs that were previously carried out by humans always brings up the fear of being supplanted. We have memories of the episodes of ‘Black Mirror’ or many other productions that talk about the subject.

It is unlikely that we will be replaced by AI – at least for the moment – ​​and, although some jobs will surely disappear, most will be transformed to include new technologies in their processes.

As indicated by the BCP executive, “the role of telephone operators will migrate towards the role of financial advisors and these professionals will need to strengthen their technological skills to manage AI systems and provide specific knowledge”.

MIRA: Four thousand password attacks per second in the world: how the misuse of artificial intelligence drives the rise of cyberattacks

The idea is that in more complex issues such as claims, theft or financial fraud, the advisor uses AI to have relevant information about the client and provide more personalized attention, so that “Not only is artificial intelligence serving people, but it also helps the agent to provide better service”.

“Technology will increase our service capacity, but the team is required to embrace this transformation and capitalize on their knowledge to offer a better customer experience”explains the general manager of IBM, who in turn points out that AI needs to be fed back.

“AI is a technology that is somehow alive and needs to be constantly trained. That’s where the knowledge and experience of the bank’s team plays a crucial role. Today, for example, we evaluate interactions throughout the day to adjust and refine use cases, achieving more precise responses. In this process, the contribution of the agents is essential”, he points out.