Amazon Chime SDKs are a set of real-time communications capabilities for developers to embed voice, video, messaging and other functions into applications.
The communications industry’s largest show, Enterprise Connect, is being held this week in Orlando, Fl. At the event, Amazon Web Services (AWS) announced its new Amazon Chime SDK call analytics, which are a set of capabilities that simplify and lower the cost of recording and generating insights into real time audio calls.
The service works with any SIP or WebRTC system, so customers can use their favorite UC or CC solution and still take advantage of the Chime SDK.
Building Communications Into Applications
Amazon Chime SDKs are a set of real-time communications capabilities for developers to embed voice, video, messaging and other functions into applications. Chime came to Amazon via the acquisition of Biba as a UC application but AWS has shifted it away from a pre-built client to the SDKs. Given how crowded the market is today, this is the right move as it addresses what I believe are the future needs of communications.
Knowledge workers are more productive with a UC client, as it facilitates better collaboration. However, other job functions would benefit from having communications integrated directly into the application.
For example, a doctor could more efficiently talk to a patient if the medical records and video were in one app instead of having to toggle between them. Also, the developer can optimize the app to the use case. An example of this is a call in a regulated vertical, which would automatically be recorded and archived instead of requiring a worker to turn on recording.
Voice Tone Analysis Extracts Sentiment
There are two core capabilities of the new SDK. The first is voice tone analysis, which uses machine learning to extract sentiment from a speech signal based on a combined analysis of lexical and linguistic information as well as acoustic and tonal data.
The product works across multiple systems and enables an organization to create a single data lake of all voice information. Also, administrators can use the AWS services to create custom dashboards to visualize the data the way they want to see it, using tools like Amazon Tableau or QuickSight.
The low hanging fruit for voice tone analysis is regulated industries where strict compliance requirements need to be met. For example, a financial firm could deploy the technology on its trading desk. The product can be used to record all the calls and use voice tone analysis to understand if traders are saying things they should not, or if the tone of a customer goes from positive to negative or vice versa as this could indicate a problem.
While industries like finance and healthcare are the most obvious, voice tone analysis can be used in every vertical as the ability to understand the sentiment of a call can be used for coaching or process improvement. The data can be analyzed, problems can be identified and corrective action taken.
Speaker Search Simplifies Caller Identification
The other capability is called speaker search, where a user is “voice printed” and the identity of that individual is stored in a database. The algorithms are very efficient and require only a small snippet of the voice to recognize a known user.
Speaker search helps expedite caller lookup and enrich call records and transcripts with identity attribution. Speaker search delivers a suggested unique internal identifier for the speaker and a confidence score. The decision to match the current speaker with a known speaker from the database can be made on an application-by-application basis depending on how it’s being used.
Bottom Line: Promoting the AWS Cloud Platform
For AWS customers, while the SDK itself adds value, there is a greater benefit. Once all the voice information is aggregated it can be stored in data lake in a location of the customers choosing. Given the customer is using AWS, it’s likely they will store in the AWS Cloud. By doing this, customers can use other AWS applications and services to get more value from the data.
As an example, Amazon Transcribe and Transcribe Call Analytics can be applied to the data lake to generate insights. Similarly, the data can be stored in an Amazon S3 bucket and voice analytics can send real-time notifications to a function deployed on AWS Lambda or an SQS queue.
Over the past few years, AWS has done an excellent job of making it easier to use a broader set of cloud services enabling customers to accelerate innovation. As communications becomes more composable, I would expect to see AWS come to market with more SDKs to enable customers to build communications-rich applications.
