Teleperformance SE, the largest call centre operator in the world, is deploying artificial intelligence across its centres in India to ‘soften’ the accents of English-speaking Indian employees, according to a Bloomberg report. The company says it would make them more ‘understandable’.
Notably, Teleperformance provides customer support to major companies like Apple, Samsung, ByteDance and TikTok.
According to the report, the technology is called accent translation and is coupled with background noise cancellation. The company’s Deputy Chief Executive Officer, Thomas Mackenbrock, said in an interview that sometimes the accent of an Indian agent is hard to understand.
He added that the technology can “neutralise the accent of the Indian speaker with zero latency,” creating “more intimacy, increases the customer satisfaction, and reduces the average handling time: it is a win-win for both parties.”
The technology was developed by a US startup Sanas. Teleperformance invested $13 million in Sanas earlier this year, in a bid to promote AI partnerships this year. It has committed to investing nearly $104 million in such partnerships this year.
While the new technology, which reduces background noise and improves accent, may be considered an aid to agents, it also carries a risk for countries like India and the Philippines, which have built market-leading positions in the sector due to their large number of English-speaking people. These new tools may diminish workers’ cultural identities and the authenticity of conversations.
According to Sanas’ website, it developed the technology to “reduce accent-based discrimination”.
It is available for Indian and Filipino accents currently, but it will soon be available for other regions, including Latin America, where Teleperformance has many workers supporting US-based customers.
Mackenbrock said that AI is ubiquitous, but the human element is ‘incredibly important’. He added that Teleperformance is also making a push to provide services to tech giants building AI models.
“Whether it is OpenAI, or any AI model in the world, it needs human support for training,” he said. This includes creating, testing and labelling data.