Every day, people call customer service centers and provide sensitive information, like credit card numbers, to agents by voice. Now, a conversational artificial intelligence (AI) solution that uses natural language understanding capabilities offers a way to remove that information from calls, while also passing data for transactions.
This is important because dealing with any type of personally identifiable information (PII) inevitably involves a series of security and privacy compliance regulations that may vary by jurisdiction. There is also a non-trivial risk that sensitive information could be leaked or stolen. In fact, there are known incidents where voice-provided credit card information has been annotated by malicious actors, leading to unintended results.
“There was an incident where a business customer approached us with a real-life story that was like, hey, look, this happened, someone wrote down credit card numbers, and that stuff leaked out onto the open market.” , Srini Bangalore, Director of AI Research at conversational AI provider Interactions, told VentureBeat. “That led us to start thinking about the technology itself and how to redact personally identifiable information in real time with low latency, without impacting the user experience.”
To that end, Interactions developed a new technology, Trustera, which is generally available as of today. The goal is to use artificial intelligence and machine learning (ML) techniques to identify PII in real time, redact it from the live voice call, and still pass the information to underlying digital systems for transactions in an encrypted approach.
Take a hybrid AI approach to conversational AI
Interactions is a company that designs conversational artificial intelligence technology platforms for organizations.
Conversational AI technology is commonly associated with human interactions with bots, but that’s not the approach that Interactions has largely taken. Bangalore said his company has taken what it called a hybrid approach to AI.
With the hybrid AI model, humans are part of the process alongside conversational AI to help support the user experience in a frictionless approach. The Trustera system, for example, is not run by bots, but is designed to work in environments where a person calls a customer service center and then talks to a human.
Bangalore said the PII redaction process in human-led conversations is more complicated than for purely digital and bot interactions in an interactive voice response (IVR) type system. He pointed out that in IVR or bot conversations, the system knows when PII is being transmitted because it is part of the process and initiated by the system.
With human-led conversations, it’s not always at the same point in a conversation that PII is requested or transferred. It is also necessary to understand what PII is being sent, as well as understanding the actual human speaker.
How Trustera’s conversational AI works to protect PII
The AI technology that Interactions has developed for its conversational AI platforms is rooted in capabilities coming from AT&T Bell Labs.
In 2014, Interactions acquired speech analytics technologies from AT&T, which is where Bangalore had previously worked for 18 years. Speech recognition capabilities have steadily improved in the years since, with the integration of Natural Language Understanding (NLU) functionality, which helps enable the Trustera service.
Interactions has trained its guard information model to understand when different human speakers transfer PII. The model is not static and is constantly updated.
“We have a self-monitoring machine learning approach, where we take calls from the previous day and have a notional confidence metric to say that these are data elements that we can add back to the model,” Bangalore said. “So we update the model periodically that way as well.”
The VentureBeat Mission is to be a digital public square for technical decision makers to learn about transformative business technology and transact. Discover our informative sessions.