BPOs used accent translators to cut costs. Its agents are paying the price
Crashing your car is bad enough. Even more so when it happens to a somewhat-superstitious American on 13 March, a Friday.
So when a call-centre representative at IT-services company Wipro, who was processing this customer’s insurance claim, inquired about the date of the accident, it was only natural for the American to joke, “Such Friday the 13th energy, am I right?”
Five seconds go by. There’s silence on the other end.
“Yes I am here,” replied the representative.
Embarrassed by the bad joke, the conversation turned slightly awkward. But in reality, the representative had laughed. It just hadn’t been captured by the AI accent translator she’d been using.
Created by Silicon Valley-based voice-AI startups such as Sanas, these translators convert on-call accents, for instance Indian or Filipino, into American ones. The reason is simple: to bridge the gap between employees’ regional accents and the polished English expected at call centres.
Companies across the board are using such translators, from health-and-insurance company Unitedhealth Group, to payment-and-banking companies like American Express and Wells Fargo, to BPOs such as Teleperformance and Alorica. They, in turn, often outsource their work to IT-services companies like Wipro, who end up using these AI accent translators by extension.
On the face of it, the tech promises great things. It frees up companies to hire cheaper talent from tier-2 and -3 towns and, in BPO parlance, improves their NPSThe ultimate scorecard for BPOs, NPS or Net Promoter Score measures customers' satisfaction with a company's products and services. For instance, Teleperformance, which used AI translators across 42,000Washington PostAI is transforming Indian call centers. What does it mean for workers? of its Indian employees, claimsTeleperformanceA consumer electronics leader delivers a 26% improvement in NPS that using Sanas helped increase its NPS by 26%.
In reality, though, the tech may translate accents but it has a hard time translating human emotives—be it a laugh, a whisper, or an inflection.
Lede illustration by Kashvi Bansal
BPOs used accent translators to cut costs. Its agents are paying the price
Related Stories
AI News
System helps humans predict when self
18 minutes ago
AI News
Hundreds protest outside Chapel Hill's G20 Innovation Ministerial, criticizing artificial intelligence
18 minutes ago
AI News
The most in
47 minutes ago
AI News
Fake 10 Downing Street listing exposes 'unfit' Booking.com, says consumer group
48 minutes ago
AI News
Most Banks Wait Six Years for AI Payback. Discovery Didn’t.
1 hour ago
AI News
US and Big Tech Advocate Looser AI Rules at G20 Gathering
2 hours ago
AI News
Anthropic Admits Security Failures Behind Claude Hacking Incidents
2 hours ago
AI News
The quiet regionalisation of AI governance
2 hours ago