VK has overhauled its recommendation algorithms for the VKontakte social network, leading to a nearly fivefold increase in views for original authors’ posts and a 22% rise in new followers.
- Recommendations now activate from the very first post, regardless of follower count.
- VK’s AI-powered machine learning models analyze post content, likelihood of subscription, and user reactions in real time.
- Engagement on new creators’ content jumped 42%, with their follower count up 54%.
The updated recommendation algorithms expose content to a broad audience immediately, bypassing the need for an existing fanbase. Along with topic and author profile data, the system analyzes the meaning of posts and estimates users’ chances of subscribing or reacting, all in real time.
This revamp also reshaped the feed logic. Original creators’ posts now feature multiples more in recommendations. These machine learning models were developed by VK’s AI engineers and are already live on VKontakte. The company plans to roll similar technology out across its other products.
”We see a clear trend: audiences are engaging more actively with original content. Users react and comment about 1.5 times more often on such posts. We’re focused on evolving recommendations to support this growth-continually refining algorithms so quality content finds the right audience faster and earns more views,” said Vera Sovetkina, head of VKontakte’s Communities and Creator Experience.
VKontakte’s approach diverges from platforms like Instagram and TikTok by enabling recommendations immediately, without requiring creators to build a baseline audience first. Its AI-driven real-time semantic analysis sets a new standard for how social networks can promote original content, especially in Russia’s ultra-competitive social media landscape, where local platforms like VK remain dominant amid global restrictions.
Going forward, it will be important to monitor how these improved recommendation algorithms affect content diversity and user retention. If VK successfully scales this system across its ecosystem, it may set a precedent for other platforms aiming to empower new creators and drive deeper engagement through smarter AI curation.