Boomerang Releases Respondable
Boomerang, an extension in both the Google Gmail and Microsoft Outlook email ecosystems, has launched Respondable, a new application of artificial intelligence that provides Boomerang users with real-time analysis of an email's quality and likeliness to garner a response, as well as suggestions for how to improve it, backed by algorithms trained on tens of millions of messages.
As emails are composed, Respondable's algorithms predict the likelihood that the message will receive a response. It also suggests improvements to increase odds of a response.
Respondable is based on the application of machine learning to a corpus of hundreds of millions of data points in public and proprietary email datasets and other datasets. Boomerang's initial research isolated a number of factors that correlate strongly with the likelihood of receiving a response to an email. They include the following:
- Message subject length;
- Message word count;
- Reading level;
- Question count;
- Politeness; and
"We believe that artificial intelligence will achieve its greatest potential working in combination with humans rather than replacing them," said Aye Moah, chief of product at Boomerang, in a statement. "Unlike applications that seek to use technology to replace humans, our philosophy is to enhance human creativity with assistance from artificial intelligence in a manner that puts humans in control."
"Building a system that acts in concert with a human as they perform creative work requires an understanding of the strengths, weaknesses, and subtleties of a variety of machine learning techniques," said Mike Chin, chief technology officer at Boomerang, in a statement. "Respondable layers several techniques together in a novel way. For example, we use deep learning and neural networks for textual analysis, but we found that other techniques worked best for assigning weights to the inscrutable outputs of neural networks in ways that can be explained to users. That allows us to make the output of our system understandable and actionable, even though the calculations are incredibly complex."
By default, Boomerang Respondable works entirely within users' mail clients. It does not transmit message data anywhere, for any purpose. If a user chooses to enable advanced machine learning features, Boomerang transmits message data using enterprise-grade encryption, analyzes the message data on a secure server, then transmits the analysis back via encrypted communication channels. Boomerang immediately discards all message data after the analysis.
Boomerang Respondable is included as part of both Boomerang for Gmail and Boomerang for Outlook. Access to Respondable for unlimited messages is included in the free Basic plan for both platforms. Advanced features, including sentiment analysis, politeness, subjectivity, and personalized analysis are available for users on Boomerang for Gmail's Pro and Premium plans.