EVO Banco: Developing a human-centric voice banking platform with Google Cloud
About EVO Banco
Founded in 2012, EVO Banco is one of Spain's largest digital banks. It is valued by over 900,000 customers for its innovative and 100% digital banking, with a focus on customer service using cutting-edge technology.
Tell us your challenge. We're here to help.
Contact us
About Altostratus
Altostratus Cloud Consulting is a Google Cloud Partner for Southern Europe, and a full-cycle cloud services provider.
EVO Banco uses AI developed on Google Cloud to create a smart automated telephone interface to solve customer queries quicker.
Google Cloud results
- Handles around 85% of calls to customer contact centers
- Comprises just 3% of the total running costs of contact centers
- Reduces average customer wait time for an answer to two minutes
- Calls processed through AI are routed accurately 95% of the time
- Results in 70% effective outcomes for customers with machine learning continuously improving this figure
First bank to develop Spanish AI voice assistance
Digital banking is booming. The COVID-19 pandemic forced banks and their customers to rapidly adopt digital tools and systems to compensate for branch, office and call center closures. Yet some banks had the foresight to predict that the future of banking was digital long before then. Among them is EVO Banco, founded in 2012 and now one of the largest digital banks in Spain.
EVO Banco's focus on innovation is what keeps it one step ahead of its competitors. It turned to Google Cloud in 2018 when the business needed help creating an AI voice banking platform for its contact center. EVO Banco had already rolled out a solution for its mobile app and website, but it needed to develop an interface that could connect the EVO virtual assistant to its customer call center to answer telephone-based queries.
The bank has been using Google Cloud for APIs since 2015, when it released its first mobile banking app, a transactional app that used Google Maps to pinpoint the locations of its ATMs. Its familiarity with the Google Cloud and its knowledge of its scope gave it the confidence to choose Google Cloud's range of products to develop its voice platform. "We needed something reliable and quick that could respond to human questions just as naturally as we are talking now," says Pedro Tomé, Head of Disruptive Innovation, Big Data and Advanced Analytics at EVO Banco.
"We didn't want a chatbot, we wanted the conversation to flow naturally as in real life. We needed a system that could connect with all the technology in our roadmap, that we could develop, scale up, and flex with demand, all while optimizing costs. That's why we chose Google Cloud."
Adapting in real-time with Google Kubernetes Engine
In the initial discussions with Google Cloud, EVO Banco was advised to bring a partner on board to help with the creation and implementation of the interface for its voice banking platform. That partner was Altostratus, a specialized consulting company belonging to Telefónica Tech. Altostratus suggested a solution based on Google Kubernetes Engine. By embedding a group of Private Branch Exchanges (PBX), based on Asterisk, within Kubernetes, calls could be quickly processed and the system adapted to call demand in real-time. "The scalability was a key feature of Kubernetes that attracted us to it," says Tomé. "It means that we can now be flexible and attend to demand as and when it happens."
Furthermore, with PBX being an open-source cloud technology that an increasing number of businesses are embracing, EVO Banco can use it to connect with any technology that a call center is using. This was important because, at the start of the project, the call center was using Avaya, but later changed to Genesys. The PBX-based solution works seamlessly with both contact center interface systems. It uses Session Initiation Protocol (SIP) to connect the different PBX technologies and a SIP trunk to connect the business PBX system to the internet to support communication.
"The contact center interface was the most important part of this project, and Altostratus helped us throughout the whole process," says Tomé. "They have really good people who are passionate about technology and experts in their roles. They were always open to supporting us in the project, even when the project had finished. They were fantastic partners to have on board."
"We didn't want a chatbot, we wanted the conversation to flow naturally as in real life. We needed a system that could connect with all the technology in our roadmap, that we could develop, scale up, and flex with demand, all while optimizing costs. That's why we chose Google Cloud."
—Pedro Tomé, Head of Disruptive Innovation, Big Data and Advanced Analytics, EVO BancoIncreasing efficiency with machine learning
As well as this interface, the EVO Banco tech stack uses API orchestration, which sees the audio from calls transcribed by the Google Cloud Speech-to-Text API in real-time. This is then sent to Dialogflow for analysis. Dialogflow understands customer intent based on this text and generates a response, whether that's answering a query or making the decision to pass more complicated queries to a human operator. All behaviors are then processed with Dataflow to develop machine-learning models that will increase the system’s efficiency over time.
"We analyzed a number of different conversational products in the market and Dialogflow was the best for us," says Tomé. "It was easy to use, offered great performance, and was also really easy to implement. Other systems we tested were much more complicated, but Dialogflow easily solved our challenges."
Creating customer-centric results
Customer experience was always the driving force behind the implementation of this technology. EVO Banco says the project was never simply about reducing person-to-person calls or replacing call-handlers with machines. It was always about delivering the best possible experience during calls, and freeing up the workforce from simple and routine inquiries to make this happen. And, what's more, the total cost of the technology accounts for just 3% of the total cost of running the customer contact centers.
"With this system, we've improved the identification of our clients to make it faster to get through security," Tomé explains. "Around 85% of the calls to the contact center are handled by Dialogflow, and the average time people spend on the line for an answer is two minutes, where it was previously five to six minutes. So, in that time, a customer has called, had a conversation with the system, obtained what they wanted and hung up."
The accuracy with which calls are correctly routed currently sits at 95%, though EVO Banco will continually improve this. And, it has developed another metric to measure not just how well calls are routed, but the quality of the outcome for its customers. "We call it conversational effectiveness, and it's currently sitting at about 70%," Tomé adds. "So that is something we are always striving to improve. To do that we measure everything and use Looker Studio and BigQuery to analyze the metrics we have. With this feedback, we can improve these conversational flows, improve the system and improve outcomes for customers."
"Around 85% of the calls to the contact center are handled by Dialogflow, and the average time people spend on the line for an answer is two minutes. So, in that time, a customer has called, had a conversation with the system, obtained what they wanted and hung up."
—Pedro Tomé, Head of Disruptive Innovation, Big Data and Advanced Analytics, EVO BancoBanking on the future with Google Cloud
"For our voice banking, we are exploring how to apply biometric technology for speaker identification, for example. As for new machine learning models, we have lots in the pipeline that I can't speak about right now. But we will be using Google's technology to make it possible."
As for what's to come from EVO Banco, Tomé says its ambitions lie in constantly improving the customer experience by trialing new technologies through Google Cloud and staying one step ahead of the competition. "EVO Banco has become a reference for digital banking in Europe and that's thanks to excellent technology combined with an excellent user experience. Our plan is to continually improve these services for our customers, using technology from Google," he concludes.
"In the financial sector we have to consider regulations and where our data is stored matters," says Tomé. "Google Cloud launching a new data center in our region not only gives us the ability to maintain our data inside our country it also allows us to leverage innovative cloud technology while complying with regulations around data residency."
As EVO Banco continues to push forward with innovation that defines the future of digital banking, Tomé says Google Cloud sits firmly at the center of the bank's machine learning plans now. The news that Google was opening a new Google data center in Spain has only helped to cement that partnership further.
"EVO Banco has become a reference for digital banking in Europe and that's thanks to excellent technology combined with an excellent user experience. Our plan is to continually improve these services for our customers, using technology from Google."
—Pedro Tomé, Head of Disruptive Innovation, Big Data and Advanced Analytics, EVO BancoTell us your challenge. We're here to help.
Contact us
About EVO Banco
Founded in 2012, EVO Banco is one of Spain's largest digital banks. It is valued by over 900,000 customers for its innovative and 100% digital banking, with a focus on customer service using cutting-edge technology.
About Altostratus
Altostratus Cloud Consulting is a Google Cloud Partner for Southern Europe, and a full-cycle cloud services provider.