Voice technologies (accent neutralization, voice cloning, empathetic AI, etc.) aim to create a seamless and “neutral” customer experience.
But by constantly smoothing out the rough edges, aren’t we at risk of stripping the relationship of its authenticity? What if, instead of reassuring people, this pursuit of technological “perfection” actually fostered mistrust?
Contents
Summary
Voice-based artificial intelligence is ushering in a new phase of transformation in customer relations by bridging the gap between technological automation and human interaction. Long associated with limited automated response systems, voice is now becoming an intelligent conversational channel capable of understanding intentions, contextualizing requests, and streamlining customer journeys.
Beyond being a technological innovation, voice AI addresses key challenges for organizations: improving service accessibility, optimizing operational performance, managing high volumes of interactions, and evolving the role of customer service representatives. It enables organizations to efficiently handle recurring requests while redirecting teams toward interactions that offer greater relational or business value.
Its deployment is thus transforming customer relationship management and contact center models. Success, however, does not depend solely on technology, but on companies’ ability to rethink their processes, governance, and performance metrics in order to integrate voice as a true strategic driver of customer experience and operational efficiency.
An Overview of New Relational Simulation Technologies
As customer interactions become increasingly digitized, new technologies for simulating customer interactions are emerging, driven by rapid advances in voice-based artificial intelligence.
These solutions help improve the quality of communication by overcoming language barriers, cultural and emotional differences, and the sometimes overly impersonal nature of automated interactions.
Several concrete solutions are already standing out on the market:
Text-to-speech with accent removal
To understand a perception (example: user satisfaction with a product)
Examples of solutions:
Sanas, Symbl.ai
Targeted irritants:
Misunderstandings due to accent, cultural biases
Maturity level:
Moderate
Use Case: Everise
Sanas integration to improve communication in healthcare call centers
Real-time translation with appropriate tone
Examples of solutions:
DeepL Voice (beta), Zoom AI Companion
Targeted Irritants:
Language barrier, emotional disconnect, loss of nuance
Maturity Level:
Low
Use Case: Deutsche Bahn
: Using DeepL to Facilitate Multilingual Communication Within the Company
Vocal production expressively enhanced by AI
Main Feature:
Generate text or voice messages with a (empathy, enthusiasm, calmness, etc.)
Examples of solutions:
GPT-4 + ElevenLabs, WellSaid Labs, Respeecher
Targeted irritants:
Communication perceived as cold, robotic, or monotonous
Maturity level:
Moderate
Emotional Regulation and Mitigation
Detect and mitigate negative emotions in customer interactions (voice or text), and assist agents
Uniphore, Observe.AI, Symbl.ai, Salesforce GPT
Aggression, anger, and emotional overload for staff members
Maturity level:
Moderate
Use Case: SoftBank (Japan)
— Testing an AI that modifies the voices of angry customers to make them sound calmer before they are heard by agents
Intent and Sentiment Detection
Key Feature:
Identify emotions, intentions, or subtle cues in customer interactions
, ASAPP, CallMiner, Zendesk AI
Key pain points:
Difficulty understanding the client’s actual needs, poor prioritization of requests
Maturity Level:
High
s adopted SentiSum, integrated with Zendesk, to analyze customer sentiment across various communication channels, among other things
The Blurred Line Between Improvement and Manipulation
By altering a speaker’s voice, adjusting the emotions expressed in real time, or translating speech to make it sound warmer than it originally was, these tools are redefining the norms of interaction.
For some, this ability to “smooth out” communication meets a growing need for fluidity, reassurance, and immediate understanding.
But it can also raise questions: Does the perceived authenticity of a conversation remain intact when an AI adjusts the speaker’s tone, accent, or emotion without the user being fully aware of it?
GUIDE:
Customer Relations in the Age of AI
Discover how Artificial Intelligence is transforming the field of customer relations. From personalization to automation, don’t miss this comprehensive guide.
Early real-world cases that have raised questions and even sparked controversy
This tension between performance and authenticity becomes particularly evident in early user feedback. The startup Sanas offers technology that neutralizes agents’ accents. It is used by call centers in India and the Philippines to better serve North American customers. While presented as a tool for equal opportunity, it has also drawn sharp criticism. Several researchers and academics view it as a form of “vocal whitening” or “digital whitening.” According to them, this solution erases part of the agents’ cultural identity in favor of an ideal of linguistic standardization.
Similarly, in Japan, SoftBank is testing an AI system that modifies the voices of angry customers to make them sound calmer before they are heard by agents. The system is designed to protect the mental health of call center agents, but it raises questions: Is moderating aggression a form of de-escalation… or censorship?
Even in purely internal contexts, these technologies do not always win unanimous approval. The Zoom AI Companion, which offers real-time translations and automatic meeting summaries, has been heavily criticized by users for its unreliability and even for translation errors into the wrong language. On Reddit and specialized platforms, some are already calling it a “useless” or “misleading” tool that creates distance rather than bridging it.
These technologies therefore create tension between two fundamental aspects of customer relations: efficiency and sincerity. In a context where consumers value both personalization and transparency, the use of artificially neutral or emotionally optimized voices can create a sense of dissonance. On the one hand, there is a desire to improve the customer experience; on the other, there is a risk of obscuring the true nature of the interaction—or even creating a sense of artificiality, or worse, manipulation.
Conclusion: An Ethical Challenge to Be Addressed
The ethical question therefore becomes central: Should customers be informed that they are interacting with a modified voice or a simulated emotion? Where is the line between compassionate assistance and emotional manipulation? In the coming years, companies will have to strike a balance between customer engagement and the demand for authenticity, at the risk of losing trust if perceived intentions do not align with customers’ actual expectations.
Customer Relations Department
AI at the service of customer relations: the complete guide to download
AI will profoundly transform the customer experience and greatly improve the efficiency of Customer Services. But while everyone's talking about it, few have really taken the plunge. We conducted a study
The customer relations professions of tomorrow in insurance
Customer service roles in the insurance industry must reinvent themselves. In this article, discover an overview of some inspiring initiatives. Summary: Customer service roles in the
Customer relations in the insurance industry: how are organizational models changing?
Vincent Placer, Managing Director of Teletech, shares his expert opinion on the transformation of customer relations organizations in the insurance industry. You can also download our free
Frequently Asked Questions
What Is Voice AI in Customer Service?
Voice AI refers to the use of artificial intelligence technologies to automate customer interactions via voice by interpreting requests made in natural language and providing appropriate responses or taking appropriate actions.
It generally relies on voice recognition and language understanding components to detect an intent and trigger a predefined workflow or response, with varying degrees of flexibility depending on the system.
Compared to traditional interactive voice response systems (phone menus), it offers a more fluid and less restrictive experience, even though it is still limited by predefined scenarios and processing capabilities.
Can voice AI personalize customer relationships?
Yes, voice AI makes it possible to personalize customer interactions, especially when it’s connected to CRM data: it can recognize a customer, understand the context of their request, and tailor its responses based on their history or current situation.
However, customers will only accept this personalization under certain conditions. The essential prerequisite is transparency: users must clearly understand that they are interacting with an AI, not a human. Any ambiguity on this point can lead to mistrust and undermine the experience.
Acceptance then depends on perceived value:
- quick and effective communication
- the ability to address the need seamlessly
- the option to switch to a human agent if needed
