Secure voice solutions expert Aeriandi and fraud detection pioneer Pindrop, have teamed up to make Pindrop’s technology available through Aeriandi’s cloud-based platform, amid rampant contact centre fraud – up 113% in 2017.
Oxford, UK-based Aeriandi offers a wide portfolio of solutions for telephone-based commerce, from PCI phone payments and fraud detection, through to call recording and archiving and is the only hosted provider to process over £1 billion in payments per year.
Pindrop’s “Phoneprinting” technology analyses more than 1,380 features of a call’s audio to detect the subtle anomalies that indicate fraud, and determine its device type, geo-location and carrier. The company also provides analysis of a phone number’s history and behaviour.
Tom Harwood, co-founder and Chief Product Officer at Aeriandi, said: “Contact centres are under increasing pressure to maintain security and compliance in the face of ever increasing call volumes. Offering Pindrop’s fraud protection solutions through the Aeriandi cloud-hosted platform gives our customers a simple and extremely effective way of detecting fraudulent calls, as well as verifying legitimate callers.”
Pindrop Network technology also uses machine learning to provide a pre-call layer of security and alert agents to potentially fraudulent calls before the conversation begins. The company last year found that call centre fraud rates had jumped from 1 in every 2000 calls in 2017 to 1 in every 937 in 2017, with 43% originating from mobile devices, an increase of 25%.
The partnership comes as Chinese technology giant Baidu claimed last week that it’s technology can clone a human’s voice with just seconds of training audio. Its Deep Voice programme had already managed to reproduce speaker identities with about half an hour of training data. It has now has lowered that time to 3.7 seconds.
The AI system can change a female voice to male, and a British accent to an American one. “Voice cloning is expected to have significant applications in the direction of personalization in human-machine interfaces,” the researchers write in a Baidu blog article on the study.