TensorFlow, Google’s machine-learning tool, is helping Gmail block 100 million hard-to-detect spam messages per day, Google says. 

That number represents less than 0.1% of the blocked incoming spam and is made up of the messages that “slip through,” the company adds in a blog post that encourages businesses to use TensorFlow.

The email volume blocked by TensorFlow includes “image-based messages, emails with hidden embedded content, and messages from newly created domains that try to hide a low volume of spammy messages within legitimate traffic,” Google states.

Launched in 2015,TensorFlow is an open-source machine learning (ML) framework. Its machine-learning capabilities help Gmail “identify patterns in large data sets that humans who create the rules might not catch,” the company says.

Google adds that it makes “granular decisions based on many different factors.”

In addition, machine learning “helps us personalize our spam protections to each user — what one person considers spam another person might consider an important message (think newsletter subscriptions or regular email notifications from an application),” Google adds.

Google is now experimenting with TensorFlow in other security-related areas within Gmail, including phishing and malware detection. 

The search firm reports that 1.5 billion people use Gmail every month, and 5 million paying businesses use Gmail in the workplace as a part of G Suite.