Analyzing companies' customer-facing communication and contact center interactions hold great potential. Identifying certain patterns, keywords and emotional charge of interactions can provide decision-makers with extremely valuable knowledge, but today this goldmine of information also hides a lot of risks and potential for fraud. In this article we are going to explore how TC&C is tackling the pressing issue of deepfake technology. Analyzing companies' customer-facing communication and contact center interactions hold great potential. Identifying certain patterns, keywords and emotional charge of interactions can provide decision-makers with extremely valuable knowledge, but today this goldmine of information also hides a lot of risks and potential for fraud. In this article we are going to explore how TC&C is tackling the pressing issue of deepfake technology.
Human speech alone tells a lot of information about a person without seeing the speaker’s face, thereby speech can be evaluated by several aspects. Based on our research, we aimed for an AI-based solution for legal compliance interaction recording related customer needs and applications to reach our goal: to revolutionize the quality assurance of customer service.
When calling customer services, we often hear: Your call is being recorded for so and so. However, the customer is right to ask: Why is it good for me if my calls get recorded? The answer is: bettering CX. How? By analyzing calls, measuring customers’ satisfaction levels and providing that information to contact center team leaders and decision-makers. That data is such a strong and authentic customer feedback that shortens the learning curve toward a seamless administration and a more polite customer service. Not only that, but this approach also eases the QA processes of companies/ call centers/ contact centers by offering for evaluation solely those conversations that show some kind of anomaly, such as impatience, anger, dissatisfaction, inappropriate words, or lack of certain words. One of the biggest hurdles on the use of AI is that this kind of AI-processing is threatening personal data protection. Therefore, emotion analysis is not a possible route, so we measure satisfaction by relying on exact speech elements (emphasis, pause, specific words).
The other side of this double-edged sword called AI would be malversation and fraud cases, for example fraud committed with deepfake technologies. Artificial intelligence can learn a person's voice, their tone of voice and reproduce it, even in other, foreign languages. Our clients in the compliance recording liabilities sector, e.g., banks are particularly exposed to the threat of deepfake. Deepfake technology includes not only the aforementioned voice forgery, but also photo and video forgery and even live chat. By early 2023, 37% of large US companies had experienced voice deepfake attacks and 29% video deepfake, and this is only expected to increase. What's more, the latest deepfake tools are not only easy to access but currently almost impossible to defend against. Almost…
Realizing our customers' needs, we decided to extend the CARIN QAI smart recorder analytic features with a so-called deepfake guard. CARIN Deepfake Guard is a solution that can indicate if a caller is not human, but rather just a synthesized, generated voice, a robot. According to our research this can be of great help to contact center operators, as it can help in investigating anomalies post-call, protecting personal data and finances, and the increasingly popular remote identification methods such as voice-based bank transfers.
Be it banking, insurance or customer care, online administration is convenient, yet the well-established authentication processes can no longer guarantee that your data and identity is in safe hands, but who can then? CARIN QAi and the CARIN Deepfake Guard, which is going to be first introduced to the public at the Call and Contact Centre Expo UK, in London on the 30th of November as part of a seminar session held by Csaba Juhász, CEO of TC&C and Csaba Juhász Jr, AI developer of TC&C. The exhibition participants can also get a hands-on experience at the stand of TC&C, booth number CC-N40.