Integrated Cognitive Solution for Call Centre Digital Transformation

The core capabilities of a call centre are heavily influenced by the muscle of its telecommunications strategy and with a focus on delivering modern, automated, and streamlined outcomes for its customers. Cognitive services can rapidly understand a caller’s emotional state, share it with a listening agent, and suggest better approaches to move the customer from a negative state into a more positive one.

The Scenario
A customer service call can quickly go from bad to worse if the wrong answers are given. Being able to keep track of client information, capture correct data and produce accurate information can prove very challenging.

In this case, we explore how i-Con has integrated with Azure cognitive services to enhance call centre customer satisfaction and enhance daily operations. We worked with a leading South African call centre to apply a digital transformation strategy and implement advanced cognitive services and analytics capabilities throughout the enterprise.

“Once we did our research and understood the need, being able to add the Azure components to i-Con was simple.” – Juan Araujo, CEO at ProSolutions Integration.

The Solution
Call centres that use cognitive services can extract valuable data about customer interactions. Using this data, companies can improve the customer experience by predicting churn, coaching agents to deliver a personalized experience, and identifying what steps to take to prevent customers from having to call in the first place. But what exactly are the features of a cognitive solution? And how can they benefit your call centre?

Speaker Recognition.

Accurately determine a speaker’s identity based on their unique voice characteristics, streamlining the verification process and delivering a better customer experience. Integrate voice verification into a range of call centre solutions to provide secure, frictionless customer engagements.

Maximize value from multi-speaker scenarios. Determine a speaker’s identity from within a group of enrolled speakers. The speaker recognition feature allows you to attribute speech to an individual speaker and supports multiuser voice recognition for personalized interactions.

i-Con provided a solution to analyze telephonic interactions between agents and customers. Later, the information needed to be moved into a database server to facilitate additional insights. i-Con’s speaker recognition component was able to determine whether the customer or the agent was conducting the conversation. Separating voice allowed us to transform the data into a normalized data structure.

Speech Recognition.

Even though telephony data is complex for humans to understand, today’s speech recognition technology excels at transcribing voice. Based on large volumes of telephony data, these models have the highest recognition accuracy, even in noisy environments.

Transcribing large amounts of telephony data is a common scenario for speech-to-text solutions. These systems provide stereo or mono audio in raw form with little or no post-processing. Using a speech recognition solution, a call centre can get high-quality transcriptions regardless of what system is used to capture audio.

“Leveraging Azure, the pre-built i-Con component was able to identify proper nouns in the speech” – Juan Araujo, CEO at ProSolutions Integration.

Data gathered from phone calls can be used to improve customer service, identify new marketing opportunities, and evaluate the performance of call centre agents. The business can then use transcripts to improve telemetry, identify key phrases, or analyze customer sentiment.

A solution was required whereby an audio file had to be translated into text. i-Con’s Speech Recognition component received the audio file from an SFTP server which is continuously polled for any new files. By processing the audio file, the i-Con Speech Recognition component generates a text file with all the words from the audio file. Data analysis is then performed on that information at a later stage, which was inserted into a SQL Server database.

OCR (Optical Character Recognition).

The days when call centres had to submit extensive paperwork and documentation to establish a relationship with their customers are long gone. Instead, customer onboarding and verification are handled through high-end OCR solutions. In addition, the digitization of records and texts into presentable forms helps call centres save on operating expenses and improve customer experience.

A modern AI-powered OCR solution fuses document authenticity checks with Intelligent Character Recognition, otherwise known as ICR, for document format and semantic analysis of characters where traditional OCRs do not. The improved method of performing OCR provides more opportunities for call centres to succeed.

i-Con has provided a solution whereby BEE certificate content had to be read and only certain parts of the certificate were inserted into a database to be associated with the client’s account. To accomplish this, i-Con would pick up the BEE certificates from an SFTP server and process them in real-time. It was as simple as pointing the OCR component of i-Con to the document’s path, and the output would be the content in string format with line break indicators. This also enabled selecting specific fields within the document. Once the fields had been extracted, the system could convert the data in a normalized format.

The Conclusion
Call centre systems using cognitive services can recognize the sentiment behind a caller’s words in real-time and help agents respond more effectively. An organization’s ability to convert upset customers into neutral or happy customers during a phone call is key to better customer retention. As a result, agents are able to handle a higher caseload.

As a whole, i-Con’s cognitive solutions are more skilled than humans at making complex, real-time decisions every day throughout their entire life cycle. The goal of i-Con’s cognitive solution is to automate the process of comparing a customer’s verbal application with supporting documentation, in order to provide accurate information to make an informed decision.

In the near future, cognitive computing-based customer service will become a make-or-break factor for success in a fast-paced, competitive call-centre environment.

“i-Con’s low-code approach coupled with the product’s versatility makes it simple, fast and easy to utilize any cognitive service.” – Armando Homann, Technical Director at ProSolutions Integration.

For more information about i-Con’s cognitive solutions, contact us at

more insights