Meet Aida our LLM-based moderation algorithm

Meet Aida: our LLM-based moderation algorithm. Aida recognizes nuance, learns context, and nearly doubles industry benchmarks for accuracy. An extra layer to our moderation system, Aida gives publishers time to focus on creating great content.

Let’s Talk

Meet Aida: our LLM-based moderation algorithm. Aida recognizes nuance, learns context, and nearly doubles industry benchmarks for accuracy. An extra layer to our moderation system, Aida gives publishers time to focus on creating great content.

Let’s Talk

1

Faster Decision
Making Over 10x faster, compared to industry benchmarks.

2

Better
Accuracy Up to 95% more accurate compared to industry benchmarks.

3

Understands
Nuance Aida understands publisher- specific context and nuances, and adapts as it reviews more content.

How does Aida work?

A comment is written.

It’s sent through >25 ML-powered textual models and gets a verdict.

If the textual models can’t be sure of their decision, the comment gets sent to Aida.

Aida makes a final decision— rejected comments get a written reason.

Any few remaining comments are forwarded to manual moderators.

Decision-making clarity. Because Aida is LLM-based, its interface provides more detail on the reason behind its verdict.

Want to double click into how Aida works? Read our blog.