AI-powered content moderation with Claude reduced spam by 99%

A business social network for the Americas began facing significant growth in automatically created spam company profiles.

Bots were generating registrations at scale, embedding referral links, cryptocurrency related sites, and other content that violated the platform's policies.

The growing volume of these profiles affected content quality and created an additional moderation burden.

The challenge was to stop unwanted content at its point of entry, while keeping a smooth experience for the legitimate businesses using the platform.

‍

Greencode integrated Claude directly into the platform's existing company profile creation and editing workflows.

Every time a profile is created or an existing one is edited, its content is sent to Claude to be evaluated against moderation criteria specific to the platform.

The integration uses a prompt designed for this use case and receives a structured response that the application can process programmatically to determine whether the content should be accepted or rejected.

The response goes beyond a simple classification. Claude also provides the specific reasons behind a rejection, allowing the application to use that information to give feedback to the user.

By embedding Claude directly into the existing workflow, moderation happens at the moment content is created or edited, before unwanted profiles can be published on the platform.

‍