Tech

Telus deploys AI to alter offshore call-centre agents' accents in real time

Using speech-to-speech technology from Tomato.ai, Telus Digital aims to smooth customer interactions, a strategy competitors Rogers and Bell have declined to follow

Author
Owen Mercer
Markets and Finance Editor
Published
Draft
Source: Hacker News · original
Tech
No image available
The telco giant faces swift public backlash and criticism from labour groups over the move to reduce "accent-related friction"

Telus has begun deploying artificial intelligence technology to modify the accents of its offshore call-centre agents in real time. The initiative, executed through its Telus Digital unit, utilizes a speech-to-speech tool supplied by the vendor Tomato.ai. The company states that the primary objective of this measure is to reduce accent-related friction for customers during their interactions with support staff.

This deployment marks a significant shift in how customer-facing services are being delivered, as the system processes audio to alter phonetic patterns instantly. While Telus frames the move as a customer service improvement, the strategy has triggered immediate and swift public backlash across Canada. The rapid reaction highlights growing scrutiny regarding the use of real-time voice conversion in live contact-centre environments.

Labour groups have been vocal in their opposition to the practice, describing the alteration of agents' voices as deceptive. Advocates are now calling for mandatory disclosure to ensure customers are aware they are speaking to an offshore agent whose accent has been digitally modified. These concerns centre on transparency and the potential implications for worker rights within the telecommunications sector.

In contrast to Telus's approach, major competitors Rogers and Bell have publicly stated they have no plans to adopt similar voice-altering technology. Both firms confirmed to The Globe and Mail that they do not intend to deploy systems that modify the accents of their support staff. This divergence in strategy suggests that while the technology is available, its adoption within the broader telecom market remains selective and contentious.

The technical implementation involves complex trade-offs between latency, naturalness, and robustness against background noise. Real-time speech-to-speech systems typically combine automatic speech recognition with neural vocoders to mask speaker identity and smooth prosody. However, integrating such pipelines into live contact-centre stacks presents operational challenges that must be balanced against the goal of reducing perceived accent differences.

As the rollout continues, the industry faces a broader conversation about consent and transparency in customer-facing services. The decision by Telus to proceed despite the criticism and the refusal of rivals to follow suit may influence how broadly this approach is adopted in the future. The situation underscores the tension between operational efficiency and the ethical considerations surrounding digital voice manipulation.

Continue reading

More from Tech

Read next: Apple to roll out manual EQ controls for AirPods in iOS 27 update
Read next: Apple rolls out visionOS 27, integrating AI-driven Siri into Vision Pro headset
Read next: Apple Overhauls Siri with Google Gemini Partnership and Standalone App at WWDC 2026