Before Deepfakes Were a Problem, Auraya Was Building the Defence
Auraya | Consensus Award Winner 2013
Today, hearing someone's voice is no longer proof that you're actually talking to them.
Artificial intelligence can imitate voices. Fraudsters can manipulate recordings. Deepfakes have transformed identity fraud from a futuristic concern into a very real security problem.
But more than a decade ago, an Australian technology company was already asking a fundamental question:
How can a computer know that you really are who you sound like?
The company was Auraya.
And its technology was ArmorVox.
In 2013, ArmorVox was recognised with a Consensus Award.
Years before AI-generated voices became part of everyday conversation, Auraya was developing technology capable of using the human voice itself as a form of identity.
Your Voice as Your Password
Voice biometrics works very differently from conventional speech recognition.
The technology isn't primarily interested in what someone says.
It analyses characteristics of who is saying it.
ArmorVox creates a voiceprint based on the acoustic and behavioural characteristics contained within speech and uses it to verify a person's identity.
That opened up significant possibilities for organisations where proving the identity of someone at the other end of a telephone or digital interaction was critical.
Banks.
Government agencies.
Healthcare organisations.
Contact centres.
Financial services.
And as digital interactions expanded, the problem Auraya was solving only became more important.
Australian Intellectual Property Goes International
Auraya continued developing the technology after its Consensus recognition.
In August 2017, the company announced it had received patent recognition in China for fraud-prevention technology using voice biometrics.
The innovation enabled organisations to combine voice verification with randomly generated challenges — helping establish both that the correct person was speaking and that they were responding to the challenge presented to them.
Auraya subsequently documented deployments and case studies involving organisations including the Bank of New Zealand, Telstra, NSW Health, OFX and New Zealand's Inland Revenue Department.
Then the technology landscape changed dramatically.
Generative AI arrived.
Suddenly, protecting someone's identity from synthetic and manipulated voices became an entirely different challenge.
And ArmorVox was still there.
More Than a Decade Later — ArmorVox 18
In November 2024, Auraya released ArmorVox 18.
More than eleven years after its Consensus recognition, the Australian-developed technology was still being actively advanced for a new generation of identity threats.
Auraya has since expanded its work around fraud detection and protection against deepfake voice attacks, while integrating its technology into major contact-centre ecosystems.
In 2025, the company announced further expansion through platforms including Five9 and NICE CXexchange.
The problem ArmorVox was originally built to address hadn't disappeared.
The world had caught up with it.
Being Early Can Feel a Lot Like Being Wrong
There's an important lesson here for innovators. Some technologies solve problems everyone already understands. Others solve problems whose importance won't become obvious for years. Building the latter can be incredibly difficult.
Customers need educating. Markets need developing. And founders can spend years waiting for technology, behaviour or regulation to catch up with what they've already seen coming.
Auraya's journey demonstrates the value of staying with a problem.
Because in 2013, voice biometrics could easily have sounded like specialist security technology.
A decade later, in a world capable of cloning a person's voice using artificial intelligence, the question of whether you can prove who is actually speaking has become considerably more important.