Telus Deploys Real‑Time Voice Conversion to Reduce Accent Friction in Call Centers
Check your contact‑centre AI pipeline for latency and naturalness trade‑offs when adding real‑time voice conversion.
Review your contact‑centre AI strategy for compliance with disclosure and worker‑rights regulations.
Summary
Telus is using a real‑time speech‑to‑speech system from Tomato.ai to modify offshore call‑centre agents’ accents, aiming to reduce “accent‑related friction” for Canadian customers.
The system runs live in the call‑centre stack, converting the agent’s voice before it reaches the caller. Telus claims the technology improves customer experience, but labour groups argue it is deceptive and call for mandatory disclosure. The Globe and Mail reports that Rogers and Bell have said they will not adopt similar voice‑altering tools. Public backlash in Canada has already erupted, with regulators urged to require disclosure of such AI‑driven voice changes. The deployment highlights trade‑offs between latency, naturalness, and robustness to background noise in real‑time voice conversion. It also raises worker‑rights concerns, as offshore agents are not notified that their accent is being altered. The story underscores the need for transparency policies and technical audits in contact‑centre AI pipelines.
Key changes
- Telus uses Tomato.ai real‑time speech‑to‑speech to alter offshore agents’ accents.
- The system runs live in the call‑centre stack, converting the agent’s voice before it reaches the caller.
- Telus claims it reduces “accent‑related friction” for customers.
- Labour groups criticize the practice as deceptive and call for mandatory disclosure.
- Rogers and Bell have said they will not adopt similar voice‑altering tools.
- Public backlash in Canada has erupted, with regulators urged to require disclosure.
- The deployment highlights worker‑rights concerns and the need for transparency policies.