In May 2025, AWS announced it would end support for its voice biometrics solution: Voice ID.
Two months later, Google removed all references to Speaker ID from its website while redirecting visitors from its voice biometric solution’s web page.
Meanwhile, Microsoft is slowly retreating from the space. Its Nuance Gatekeeper offering once dominated the market. However, its share dropped as the original Nuance team departed, and the institutional knowledge has fallen away.
Matt Smallman, Founder of SymNex Consulting, made this observation in a conversation with CX Today. He said:
They’re still making iterative improvements, but relationships with major clients have faded, the roadmap is unclear, and the product sits outside Microsoft’s core tech stack… I’m 95 percent sure the product will look very different, or be gone, in five years.
Some industry observers have claimed the product has been off-sale since last year. However, Microsoft hasn’t made an official announcement.
Why Are Tech Giants Abandoning Voice Biometrics?
Voice biometrics for call centers has always been a niche market. Indeed, many businesses don’t implement it due to concerns around costs and abiding by privacy laws.
Therefore, it’s a marginal product for all three tech giants, even Microsoft, which owned 80-90 percent of the market when it acquired Nuance in 2022.
Yet, not only is it a marginal product, it doesn’t quite fit their operating models, especially Microsoft’s, which doesn’t favor the high-touch, professional services approach required for a successful deployment.
Regarding Microsoft specifically, it has reduced its voice biometrics product team, as noted earlier. It’s also rolled most sales into its Dynamics Contact Center 365 platform, and the solution hasn’t had the focus it once did.
Smallman noted: “My view is they didn’t make a deliberate exit decision, but a series of decisions that will likely lead to the product fading over time.”
Regarding AWS and Google, Smallman noted: “They produced ‘good enough’ products at unbeatable price points but without the customization, fraud watchlists, or synthetic speech detection needed by top-tier banks.
They never achieved deep product-market fit, and as synthetic speech becomes harder to detect, the reputational risk of a compromise grew, making it easier for them to step away.
Who Are the Main Players Now?
Pindrop is perhaps the most prominent provider to keep pace with the latest industry advances. Meanwhile, Daon and Veridas also offer voice biometrics as part of broader identity solutions.
Smaller niche players that are innovating in this space include Auraya (Australia), ValidSoft (US/UK), Voice Biometrics Group (US), and InGenID (US).
Some CCaaS providers also offer voice biometrics, but typically sell through partnerships. For example, Talkdesk has a voice biometric solution, but it’s white-labeled from another vendor.
There is also Omilia, which focuses on voice self-service, and Phonexia from Czechia, which works in public safety and national security. These also offer robust solutions, per Smallman.
Yet, the consultant summarized:
The market is still in flux because Nuance dominated for so long. As contracts expire, others are positioning as alternatives, seeking investment, and enhancing their offerings.
The Future of Voice Biometrics in the Age of Deepfakes
To Smallman’s earlier point, synthetic speech is becoming harder to track. As this technology has evolved, so has chatter around fraudsters being able to breach voice biometrics systems.
Go back to April 2024, and OpenAI warned businesses to take steps “like phasing out voice-based authentication” ahead of the launch of its voice cloning tool. 16 months later, and the AI giant is still delaying that solution’s release.
Nevertheless, the “main players” above are taking measures to stay one step ahead, such as tracking telltale artifacts. Moreover, voice biometrics will always be powerful as part of a multi-factor authentication system.
Yet, there’s more that providers can do to bolster their authentication systems, so long as telcos play ball. As Smallman said:
The real missed opportunity is telcos not exposing device-level authentication data to enterprises, which could massively improve call security.
However, the biggest risk when it comes to voice fraud is synthetic speech in scams targeting vulnerable people.
For instance, an attacker could call a customer, convince them they’re an employee, and trick them into revealing credentials or authorizing payments.
While that risk affects knowledge-based credentials far more than biometrics, it’s the biggest threat deepfakes currently pose to customer safety.
