Krisp upgrades multilingual voice AI translation for call centers
CALIFORNIA, UNITED STATES — Krisp, an AI-powered software, has unveiled a major upgrade to its AI voice translation technology, enabling seamless, real-time multilingual conversations.
According to the company’s press release, the new Synchronous mode allows callers to speak and hear in their native languages simultaneously, eliminating traditional interpretation delays.
Real-time multilingual conversation without delays
The technology represents a leap in technology, such that it relies on state-of-the-art AI and deep learning to feature immediate bidirectional speech-to-speech translation capability, negating the need to employ third-party language service providers and the inherent scheduling, time-related, and staffing challenges that might be encountered in the process of organizing such a process
The contribution of this innovation can be measured, and it is high. According to Krisp, the Synchronous mode achieves 5 to 10 more Bilingual Evaluation Understudy (BLEU) points of translation accuracy when compared to the previous mode.
This addition makes the process of conversing seem less awkward as a direct translation but rather more like the normal back-and-forth of an oral exchange, to the point of being able to handle challenging scenarios that might include figures, IDs, and other personally identifiable information (PII), all without loss to the emotional context that the speakers originally worked with.
“From instant, natural-sounding translation to automated quality checks and effortless deployment, we’re transforming how organizations connect with customers across languages,” said Davit Baghdasaryan, Chief Executive Officer (CEO) and Co-founder of Krisp.
“Beyond a technology upgrade, this tool drives faster resolutions and stronger customer relationships to improve performance at every level of the contact center.”
Automated quality scoring for enterprise calls
Beyond core translation, Krisp’s update is focused on providing enterprises with robust tools for quality control, branding, and security. The auto-scoring feature will automatically score 100 percent of translated calls, meaning it will dispense with cumbersome and imprecise manual sampling, providing uniform quality checks using data-driven analysis to support compliance and service levels.
This is further enhanced by custom prompts, which can be used to pre-record branded and language-specific opening messages that set clear expectations and give the company a unique tone of voice at a very early stage of an interaction.
This solution is designed to work across the world and meet highly advanced operational requirements. It is multilingual and supports more than 80 languages with critical features such as language auto-detection, which identifies the language of a customer within seconds to activate the translation process, and a customer vocabulary allowing the loading of industry-specific words to ensure a higher rate of accuracy.
Built with enterprise-level management and demanding security requirements, the platform can be deployed confidently across distributed teams and high-volume operations in contact centers.
Enhancing the call center experience
This technology directly addresses two of the most critical elements for customer satisfaction: personalization and speed. It puts customers in control of the experience because the tool identifies the language that a caller is speaking immediately and, thus, allows them to express it in their natural way.
This reduces the time it takes to solve a customer problem, enhances customer rapport, and makes the call center more efficient because language issues do not hold it back, and delivers improved performance across the board.
