Trustera is the first functional system that redacts personally identifiable information (PII) in real-time spoken conversations. This removes agents’ need to hear sensitive information while preserving the naturalness of live customer-agent conversations.

As opposed to post-call redaction, audio masking starts as soon as the customer begins speaking to a PII entity. This significantly reduces the risk of PII being intercepted or stored in insecure data storage. Trustera's architecture consists of a pipeline of automatic speech recognition, natural language understanding, and a live audio redactor module. The system's goal is three-fold: redact entities that are PII, mask the audio that goes to the agent, and at the same time capture the entity, so that the captured PII can be used for a payment transaction or caller identification. Trustera is currently being used by thousands of agents to secure customers' sensitive information.

Trustera’s goal is to (1) identify PII entities, (2) mask their mention in the audio and transcripts, and (3) capture the entity values for secure transaction processing. Figure 1 demonstrates secure credit card number redaction by Trustera. We describe an early lab version of our solution in the paragraphs below.

Incoming stereo audio is decoded with two ASR decoders for the agent and caller. Live audio redactor (LAR) monitors the partial hypotheses and it triggers the NLU module as soon as a digit is detected in the hypothesis. NLU predicts the type of the entity and it informs LAR if the coming entity needs to be redacted. Then, LAR masks the audio till the end of entity is decided. The canonical value of the entity is extracted and sent to the payment system to complete the transaction.