RAG & Search
Collaborative filtering
Making suggestions based on what people with similar interests or habits have chosen.
Example
A service suggests films enjoyed by viewers with similar viewing histories.
Why people use it
It suggests things using patterns in what other people have chosen.
What you'll hear
“People with similar tastes liked this too.”
What this means for you
Be aware that sparse histories and popularity effects can shape recommendations.
Can you control it?
Developer-only
The people building or running the AI choose this setup. An everyday user generally needs their help to change how this part works.
Common questions
- Does it require knowing what a film is about?
- Not necessarily. It can work from interaction patterns alone.
- Why can a new user get weak suggestions?
- The service may not yet have enough information about that person's choices.
- Can popular items dominate?
- Yes. Items with many interactions can receive more attention than unfamiliar options.