An overview of Resource Approvals

When you log into your Ariglad environment you'll see a list of suggested additions and changes to your knowledge base. There are two types of changes: a 'New' article, where the AI has created a completely new resource and an 'Update' which is suggesting changes to an existing article. When you approve these suggestions, they will automatically be pushed out into your environment.

Team members can also click ‘Re-write with AI’ if they would like the new copy to be modified in how it’s written, for example if they want it to be explained differently or change its tone. Once a change is approved, the copy will be automatically pushed to your knowledge base. When the new copy integrates with existing articles, it will replicate the formatting of that article

How suggestions are made

After analyzing vast amounts of communication data between agents and customers, and cross-referencing the data with the knowledge base, Ariglad AI identifies knowledge gaps and builds the foundation to fill them by:
- Identifying common questions or topics raised by customers that are not comprehensively covered in the knowledge base
- Pin-pointing gaps in existing articles that excludes information commonly provided to customers by CSRs

When information gaps are identified in the knowledge base, Ariglad uses data provided to customers by CSRs to craft additions to the KB from scratch. Ariglad mimics the tone of voice that CSRs use when speaking to customers and matches that with the formatting and tone of the existing KB language. Additions to the KB can range from a sentence addition to a paragraph, or a completely new article if there isn’t one already covering that topic or question

Discarding approvals

If you see a suggestion that you don't want to push to your knowledge base, you can easily discard the suggestion by clicking the 'Discard' button on the top right corner of the suggestion page. When you discard or approve articles, you are letting the AI algorithm know which articles you found/didn't find helpful. This helps the AI provide you with better proposed changes moving forward .