Community content does more than inform; it builds knowledge, drives engagement, and connects members to each other. The Enrichment Toolkit applies AI to make that content easier to find, easier to read, and better organised, without adding work for authors or administrators.

Semantic search

The toolkit's most powerful capability is semantic search: the ability to find content by meaning rather than by matching exact words. A member searching for "how do I reset my account" will surface threads titled "I can't log in" — because the model understands they are asking the same question.

Two backend options are provided, so you can use the vector store that fits your infrastructure.

Auto tag suggestion

When a member creates a thread without adding any tags, a widget intercepts the submission, sends the draft body text to the Contextuly API, and returns a ranked list of suggested tags. A confirmation modal shows the suggestions to the author, who can accept all, select individual tags, or dismiss entirely.

Authors remain in control. Tags are never added without confirmation.

Configuration options for administrators:

• Minimum confidence threshold — only surface suggestions above a set score
• Preferred tag list — tags to promote in suggestions
• Excluded tag list — tags to suppress entirely
• Keyword-to-tag rules — map specific keywords in content to specific tags

This increases tag completeness across the community without requiring authors to remember your taxonomy.

Thread and blog summarisation

Long threads are valuable, but they take time to read. The summarisation features give members an accurate AI-generated summary of a discussion thread or blog post, so they can decide whether to read in full before committing the time.

• Thread summarisation — installs a widget on discussion threads that displays an AI-generated summary. Configurable via the widget API for placement and styling.
• Blog post summary — at create or update time, the plugin generates an accurate summary of the blog post body and stores it as the excerpt. This replaces the default behaviour of auto-truncating the first paragraph, which rarely produces a useful summary.

Alt Text Auto caption

Uploaded images without alt text are inaccessible to screen reader users and invisible to search. The Auto Caption plugin uses AI to generate descriptive alt text for images at upload time.

• Improves accessibility for members using screen readers, with no action required from the author
• Provides more accurate image context for search indexing
• Requires no change to the authoring workflow

Your data stays private

All AI in the Enrichment Toolkit is processed on infrastructure owned and operated by 4 Roads. Community content is never sent to consumer AI services such as OpenAI, Google AI, or Anthropic.

What it means for your community

For members: search that understands what they mean, not just what they type. Thread summaries that help them navigate long discussions. Tags that make their content discoverable without extra effort.

For community managers: a more organised, more searchable knowledge base that grows with the community. Tagging becomes a default behaviour rather than something you have to chase authors to do.

For platform administrators: all AI capability without any dependency on third-party AI providers. No data leaves your infrastructure during semantic search. No contracts with OpenAI or Google required.

Requirements

• Verint Community 12 or later
• Semantic Search requires a MongoDB Atlas cluster (M10 or higher) or an Azure Cosmos DB for NoSQL account with vector search enabled, plus a Contextuly API subscription
• Auto Tag Suggestion, Thread Summarisation, Blog Post Summary, and Auto Caption require a Contextuly API subscription (Part of 4 Roads)