Big Data mining for better contact centre performance
We’re all familiar with this pre-recorded, often robotic voice that tells us how phone conversations with the likes of e-commerce businesses or after-sale services “may be recorded for quality assurance purposes.” Now if you ever wondered what these companies were actually doing with the recordings… the truth is, not as much as they could. To date, contact centres have only been able to analyse a fraction of the calls they record. They often do this manually or with rudimentary software, and undoubtedly miss out on very important trends and issues in the process. The BISON (BIg Speech data analytics for cONtact centres) project hoped to toss this problem onto the garbage heap of history with the help of innovative Big Data mining software. Their solution is already being used in several contact centres in Central Europe, and they have big plans for the rest of the continent. What would industry stand to gain from more evolved use of contact centre data mining? Mr Marek Klimes: Better efficiency and consequently lower costs. Nowadays, contact centres are able to listen to 1-3 % of calls only. With contact centre data mining, you would be able to get information from 100 % of calls to support your decisions. Who should be trained in your team, what do your customers want or what are the emerging topics? All this information is available in your calls. What are the most innovative aspects of your approach to such data mining? We have used the complete portfolio of speech technologies, including both speech analytics technologies (speech transcription and keyword spotting) and voice biometrics in multiple European languages. This enabled our teams to cover diverse use cases in contact centres. Besides its compliance with the law, our product also takes into account new challenges related to Big Data and the anonymisation of private data. What legal aspects did you focus on and why? Legal and ethical aspects were often perceived as an obstacle to the creation of a good product. The BISON consortium believed in the opportunity to create a product complying with all necessary legal requirements by design. We have created the BISON societal and ethical code, which helps potential BISON users from the earliest deployment to actual usage. This code answers questions from four main pillars: How EU data protection rules can effectively protect citizens from modern technologies; how the EU regulatory framework can be exploited to develop law-abiding technologies; how to develop an ethical system respecting user privacy; and how BISON handles privacy issues. Can you provide an example of possibilities brought by your technology? There are many different ways to leverage the solution developed under the BISON project. Broadly speaking, you can unveil blind spots in contact centres that result in higher costs. To be more specific, we can tell you if contact centre agents speak too fast, interrupt customers or are having overly long monologues or even what the most common topic brought up during calls is. Another example: if you are suddenly handling more calls about a problem with internet connections, you will know it from our topic detection tool. Finally, we provide long-term statistics displaying the progress of contact centres in easy-to-understand graphic layers displayed in the BISON dashboard. What were the main outcomes of prototype testing? We confirmed the soundness of our vision for the contact centre market and gathered valuable user feedback. At the same time, we also found out about various missing details in the prototype. Direct connections with contact centres within the BISON consortium helped us improve our reporting tools and usage of the BISON recording management tool. What has been the feedback from industry so far? We have had positive reactions on how our system deals with unstructured data. The problem currently being faced by contact centres is not so much the lack of data but rather the lack of solutions to learn from. A typical contact centre produces a wealth of multilingual spoken data that is nowadays mined by humans or by rudimentary technical means. Our system automates this process. What are your plans for commercialisation? From the very start of the project, we have been striving for the commercialisation of BISON. Due to the geographical position of project participants and the 14 European languages covered in the project, we focus on Central Europe with the plan to spread our system across the rest of the continent. I am glad to announce that we have already successfully deployed our system in production environments across several call centres.
Countries
Czechia
