AI has exploded in the CCaaS sphere in a move to bring long-coveted efficiency to contact centers amid demanding customer expectations.
Now, the average contact center agent has everything from in-call AI prompts to help agents better answer customer queries to AI-assisted handovers aiding understanding from call to call.
Indeed, it seems that every AI efficiency that can be applied to a contact center is being applied.
But far from increasing agent performance on a call or using analytics to reduce bottlenecks, generally well-received by customers who spend less time on hold or repeating themselves after transfers, one of its applications in a contact center is splitting opinion.
AI has reached a level where companies are now able to use it to change the accent of their contact center agents.
“Accent Conversion fundamentally shifts how contact centers support their agents,”
CEO and Co-Founder Davit Baghdasaryan told UC Today.
Although proponents espouse the benefits it brings to contact center operations, critics cry foul over AI erasing agents’ personalities and enforcing a culture of a “correct accent.”
Thus, a debate has opened up for potential customers regarding these AI accent solutions: should you implement an AI technology that could cause a big customer backlash?
When AI Learns to Speak Like You
Contact centers are increasingly located in areas where English is not the main language.
Indeed, Mordor Intelligence found that Asia-Pacific is the fastest-growing region for contact center presence, driven mostly by the cost advantages it brings.
Equally, call center agent turnover rates exceed 30% annually, according to contact center platform Squaretalk.
That churn means companies using contact centers in these areas often hire workers based on need rather than attributes.
In many cases, this can result in workers whose comprehension of English is not as well established as those in contact centers in English-speaking countries.
Therefore, as a way address this, companies like Krisp have developed AI solution to keep the contact center agents but change their accents.
“For contact centers in high-friction regions, there is a preference to implement accent-changing technology, as strong accents can cause consistent communication barriers, negatively impacting metrics like first-call resolution (FCR) or customer satisfaction (CSAT),”
Sandeep Marwah, vice president, CCaaS consulting at HGS, told UC Today.
AI accent technology works by analyzing an agent’s speech patterns and converting their natural accent to what is typically described as a “neutral” or “standard” accent, often resembling American English pronunciation.
“Accent-changing technologies can streamline communications in markets with significant linguistic diversity, potentially improving average handle time (AHT). They can also enhance clarity and reduce misunderstandings in customer interactions,” Marwah said.
This is what Baghdasaryan explains was his inspiration for creating the solution.
“As a non-native English speaker myself, I’ve experienced the friction that comes with being misunderstood – not because of what I’m saying, but how I sound,”
he said.
“Every day, tens of millions of people face this same challenge. It’s not just inconvenient, it’s emotionally exhausting and can be professionally limiting.”
Proponents argue this innovation addresses genuine communication challenges that affect millions of contact center workers globally. At the same time, critics view it as a digital form of cultural erasure that reinforces harmful biases about whose voices are considered “acceptable” in professional settings.
AI Accent-Changing: A Reputation Risk?
As early as 2022, early adoptions of AI accent-changing technology created by companies like Sanas were accused of trying to make the world sound whiter by the seventh most popular news site in the US, SFGATE.
These types of sentiments are not isolated, and the internet is littered with articles from various publications, from The Guardian to Business Insider, discussing this issue.
Therefore, companies that use these technologies in their contact centers could face serious backlash from customers who consider them insensitive.
“Potential drawbacks of using the technology include diminished brand authenticity, regulatory scrutiny around voice deception, and potential negative impact on net promoter score (NPS) due to perceived lack of transparency,”
Marwah said.
“The use of synthetic voice modulation also raises questions around digital trust, agent identity obfuscation, and inclusivity optics. These are increasingly critical in regions governed by data ethics frameworks like GDPR and POPIA.”
Already, discussions around dissatisfaction with AI have seen consumer backlash. Google apologized for ‘missing the mark’ after its Gemini AI image creation tool created historically inaccurate images in response to prompts to create pictures of figures key to America’s founding.
Thus, in a move to save costs by improving call efficiency with AI accent-changing technology, companies could stand to lose money as a result of any reputational damage.
Losing Their Voice or Leveling the Playing Field?
Supporters of accent modification technology frame the debate differently, arguing that the technology serves as an accessibility tool that levels the playing field for non-native speakers.
They contend that the real discrimination lies in allowing accent-based bias to continue affecting career opportunities and customer interactions.
“Fast and effective communication is critical in call centers, where gaps in understanding can disrupt service, extend call times, and place excessive pressure on agents,” Baghdasaryan said.
“These challenges often stem not from how agents speak, but from how they’re perceived and understood by customers. The issue isn’t speech, it’s comprehension.”
From this perspective, accent conversion technology shifts the burden of addressing communication barriers from individual agents to technological solutions.
Rather than requiring agents to spend months in accent training programs to achieve this, the technology provides immediate support while preserving agents’ ability to communicate naturally and lean on their other skills not tied to their accent to progress in their jobs.
“By addressing accent-related bias in hiring, organizations can also build more inclusive teams without compromising communication standards and service quality.”
The operational benefits, according to Baghdasaryan, are significant. Krisp reports that companies using the technology have seen improvements across the board: faster resolutions, reduced handle times, higher customer satisfaction, and better agent retention.
Equally, Baghdasaryan argues that rather than being an unwelcome imposition on agents, Krisp’s feedback from them has been notably positive.
Many are reporting reduced stress and increased confidence during customer interactions, allowing agents to focus on problem-solving rather than constantly monitoring how their accent might be perceived, potentially leading to more effective customer service.
Instead of eliminating a person’s personality, Baghdasaryan describes it as “a real-time, AI-powered experience that improves comprehension and human connection.”
Marwah agrees that agent comprehension is important; however, he advocates a more organic approach to addressing comprehension issues while balancing customers’ concerns.
“In today’s customer-centric era, each and every customer increasingly values authentic, culturally diverse interactions,”
he said.
“Over-engineering accents might backfire by making interactions feel less genuine. While accent-modifying technology is technically advancing, most enterprise-grade CCaaS operations prioritize genuine agent-customer rapport, linguistic adaptability, and cultural fluency over artificial modulation.”
As a result, Marwah says HGS advocates choosing locations for contact centers that do not suffer from these comprehension issues but still offer the benefits of offshoring.
“At HGS, we have identified South Africa as a strategic CX and business process outsourcing (BPO) delivery location. The nation is globally recognized for its neutral English accent, high linguistic proficiency, and cultural alignment with Western markets.
“The need for accent modification is minimized due to the phonetic clarity and intelligibility of South African agents, which supports high voice channel performance metrics without compromising authenticity.”
Making the Right Choice on Voice
The future of accent modification technology likely depends on how companies navigate the tension between operational efficiency and cultural sensitivity.
For Baghdasaryan, explaining his and Krisp’s vision for AI accent-changing technology often assuages any doubt companies may have.
“When we connect with potential prospects with these concerns, it’s critical to emphasize that the technology doesn’t change who an individual is; its purpose is to ensure agents are understood and mitigate any communication barriers, stereotypes, and bias that unfairly impact performance,” he said.
One approach that there is a consensus on between Baghdasaryan and Marwah is that transparency is key.
“We advise clients to clearly disclose Accent Conversion, either through IVR or opt-in options during calls, as a tool that improves communication and quality of service without impacting identity,”
Baghdasaryan explained.
Ultimately, the adoption of accent modification technology hinges on how organizations balance operational gains with cultural and ethical considerations.
But for now, the question isn’t just whether we can change accents with AI, but whether doing so will help or hinder operations.
