Amazon Transcribe now supports automatic content redaction

Posted on: Feb 26, 2020

Amazon Transcribe is an automatic speech recognition service that you can use to add speech-to-text capability to your applications. Starting today, you can instruct Amazon Transcribe to automatically remove sensitive personally identifiable information (PII) from your transcription results. Amazon Transcribe uses state-of-the-art machine learning technology to accurately identify and remove PII. PII removed includes social security number, credit card/bank account information and contact information such as name, email address, phone number and mailing address. PII redaction enables contact centers to review and share transcripts to improve the customer service experience, coach agents, and discover new business opportunities while maintaining their standards for handling PII. 

Contact centers often rely on an agent remembering to pause call recording when collecting sensitive personal information. This approach is error prone and distracts the agent from having a focused conversation with the customer. There are solutions that exist for this problem. However, most of these solutions use a rule-based method that is overly simplistic and removes any number longer than two digits causing over redaction. These solutions also fail to redact non-numeric PII such as names, email addresses and mailing addresses. Other solutions that offer redaction of non-numeric personal information expect well formatted text input and perform poorly on PII in conversational form. Amazon Transcribe’s redaction feature is purpose-built for identifying PII in conversational transcripts and is highly accurate for both numeric and non-numeric PII.

Automatic content redaction is available now for US English with asynchronous transcription jobs in the following AWS regions: US East (N. Virginia), US West (N. California), US East (Ohio), US West (Oregon), Asia Pacific (Mumbai), Asia Pacific (Seoul), Asia Pacific (Singapore), Asia Pacific (Sydney), Asia Pacific (Hong Kong), Asia Pacific (Tokyo), Canada (Central), EU (Frankfurt), EU (Ireland), EU (London), EU (Paris), South America (Sao Paulo), Middle East (Bahrain), and AWS GovCloud (US-West). You will incur additional charges as described in Automatic content redaction pricing. To learn more, see the AWS News Blog post and Amazon Transcribe documentation