“Automated Anomaly Detection for Call Center” is a solution developed by FPT Software using new technologies like AI, IoT, Cloud Computing, Big Data… As per its name, the solution is for detecting anomalies in calls from customers, analyzes call quality, and optimize business operations, customer experience, as well as satisfaction.

Investing in customer care is starting to become a trend of the business world. In particular, companies are rushing to find any methods for the best customer experience, from mail marketing, sending messages, to building advertising websites and so on. Among those, call center is one attractive channel to many businesses.

While companies are pushing customers to contact them in other ways, telephone remains the most popular channel when they require service. According to Forrester, call center takes up 73% of customer contacts, and the number of complaints sent via calls are 11 times higher compares to via websites.

Currently, call center is the norm, but the question is, how to make a difference? How to improve the quality of call centers, making it the ace to advantages and building a robust and reliable relationship with potential customers?

FWI Division from FPT Software has recently built and develop a successful AI solution in automatizing call quality control for call centers. As supervision and detection of abnormal calls are usually done via recorded sound files, or integrated into the call center system, the solution is expected to be extremely useful in its assistance of customer support, answering to all their inquiries using the most effective way.

The more the challenge, the larger the will

According to Ms. Phan Thi Ngoc Nu (FWI – FPT Software), the first challenge in building this solution is having deployment data. Normally, businesses keep these data to themselves, and getting their permission for access has taken quite the great effort. Furthermore, it was also quite challenging to process data and integrate the solution to the Akaminds platform, while still ensuring the set performance.

In order to develop the solution, the project team had to process and evaluate large amounts of data. Yet, there is no standard or criteria to telling anomalies from calls, and the team needed to research, explore, evaluate and draw from experience for the final and most accurate evalutions.

Yet, it might be these very challenges that had motivated the team members to success, reaching the goal of the project.

Big benefits from efficient application of AI

There are various call center support applications on the market. However, most of them focus on pure efficiency solutions/applications to improve human performance. The “Automated Anomaly Detection for Call Center” solution, however, via its use of AI application in call center operations, does not only replace parts of human work, but also improve customer experience via its ability to analyze and understand human emotions and make decisions.

According to Dr. Nguyen Do Trung Chanh (FWI – FPT Software), the Automated Anomaly Detection for Call Centers solution can detect call anomalies via the tones of callers. First, the recording or live call will be transformed to frequency domains, which represents tone using STFT. Then, the recording or live call will be treated as an image with the diverse feature of voice. Finally, FWWI re-designed the two models GAN and Autoencoder to classify models per reconstruction damage.

Mr. Nguyen Dang Duy, a Solution Architect at FWI – FPT Software, believed that the platform design can help customers deploy new business approaches and improve current procedures at the call center’s divisions or customer care centers. He also said that, Automated Anomaly Detection for Call Centers can support a smooth, versatile integration of the solution to customers’ preexisting systems, without the need to consider any deployment infrastructure.

Main functions of the Automated Anomaly Detection for Call Center solution:

  • Integration of AI processing on call center/customer service systems: Allows deployment of machine learning models on recordings or live calls.
  • Integration of IoT data processing: Automatically processes IoT data in a flexible manner, supports deployment in a fast and clean manner, management equipment, control connection portals, data streams and pre-built development frames.
  • Easy deployment of big data analysis: Automated Anomaly Detection for Call Centers provides main components of big data analysis, including: Data lake, automatic integrated data stream, data generation, data analysis and display tools.
  • Adaptive response to customer demands: Provide standard architectures for the deployment of Automated Anomaly Detection for Call Center.

Orientation for “Automated Anomaly Detection for Call Center”

If you need a call center to process large amounts of customers with various inquiries without running into any delays or shutdowns, the “Automated Anomaly Detection for Call Center” will be the answer to all your problems. The system is constant working on a 24/7 basis as long as it has business maneuver.

The solution is also spectacular in that it can inspect and detect abnormal calls using call recording integrated to the call center system and the application of new technologies like AI, IoT, Cloud Computing, Big Data…

Businesses can adjust this solution to suit their brand and business model. There is also periodically training provided to increase the accuracy for distinctive business models.

In the near future, the project plans to re-train the solution per difference business problems of new customers. They will also adjust it to the businesses’ demands, at the same time integrating it with pre-existing call center systems, and developing more outstanding BI reports on the basis of Anomaly Detection.

In this digital age, customers are constantly moving among their various options, and constantly demanding more on service quality.

FPT’s Automatic AI solution for call quality control at call centers has applied budding technologies to detect call anomalies, with high promises to support the efficiency in knowing the customer for better service, better business performance, and added values for businesses.

FPT Software

Related posts: