Recommendations from unstructured social and chat conversations

Netflix use IV.AI's NLP platform to help their users find content to watch based on their mood and emotional state. IV.AI and Netflix created Recomoji as a way for new users and Netflix subscribers to search for content in a whole new way.

The backend emotional recommender engine was trained by 100s of millions of social conversations and driven via an automated bot published in 59 countries around the world and 32 languages including Japannese, Mandarine, French, Arabic, German, and Hebrew.

Within 24hrs the AI had processed 1 MILLION engagements with peak usages of over 250 MESSAGES PER SECOND.

Territories in 32 languages globally
Engagements in the first 24 hours
Messages received per second
More engagements vs all other social posts

By leveraging IV.AI’s language understanding platform, Netflix was able to interpret emotional input from the user and return recommendations to watch, save for later, or remove content they no longer want to see. For the first time ever Netflix users are able to curate discovery in a conversational environment so they can share with friends and drive collaborative discovery via Extensions in chat. The AI NLP learned from the way people use emojis on social media and then mapped the way they categorize their love or dislike of content, creators, and concepts so that emojis could drive discovery.

Get in touch

IV.AI makes it easy for organizations to use highly accurate AI and manage the new scale of natural language processing data. Get in touch to see the platform and meet the experts.

Francisco Brito
Chief Data Officer