Netflix does not usually jump to the top of the list when one thinks of leaders in artificial intelligence, but Netflix’s VP of Product Innovation, Carlos Uribe-Gomez, and Chief Product Officer Neil Hunt published a paper informing investors that some of its algorithms help them save $1 billion each year.

How Netflix Saves $1 Billion A Year Using AI

How the recommendation saves Netflix money

In the paper, the two executives detailed how the company’s recommendation engine impacts its churn rate. The video streaming giant does not report its churn rate, but the paper mentions that the Netflix’s retention rates “are already high enough that it takes a very meaningful improvement to make a retention difference of even 0.1%.”

This year, the streaming giant plans to spend $6 billion on content. With such a big investment, it could get all sorts of TV series and movies, but if it just presents the most popular selections to everyone, many titles would remain unseen. Not only would that mean a lot of wasted content, but it would also mean viewers would spend less time watching Netflix instead of other networks.

According to the streaming giant, the take-rate based on top personalized video recommendations was three to four times more than simply giving viewers a list of popular videos. In addition, viewers are exposed to four times as many videos using personalized recommendations compared to a list of the most popular titles.

Through these recommendations, the streaming service can even surface titles that would not usually find an audience on a traditional cable network. This allows it to save money on its content spend since it is able to increase the value of inexpensive titles.

Netflix’s AI expenses on the rise

Netflix’s technology and development budget has ramped slightly faster than its overall revenue over the past few years. It increased 38% to $651 million last year, or about 10% of revenue. Technology and development expenses include maintaining, testing and modifying the user interface, merchandising, recommendations, streaming delivery technology and infrastructure, according to Netflix’s 10-K.

In addition, the expenses also include costs associated with computer software and hardware, which includes its contract with Amazon’s AWS. Costs associated with AWS increased $23.8 million last year, and they should increase this year as well after the streaming service completes its migration to depend completely on the cloud platform.

If Uribe-Gomez and Hunt’s assessment of $1 billion per year in savings is close to accurate, then the streaming giant is seeing a significant return on its investment in AI research. As the company grows its user base, it is going to get larger.