Issue: > Currently, the Confluence article suggestions in JSM Chat and the Customer Portal are heavily dependent on exact keyword matching. When a customer types a full sentence or a complex natural language query (e.g., "I am having trouble accessing my account and need to reset my password"), the system fails to suggest relevant articles. However, if the user types only the specific document title (e.g., "Reset Password"), the suggestion works perfectly.

Impact: > This limitation reduces the self-service (deflection) rate, as customers don't usually search using exact technical terms, leading to unnecessary support tickets.

Suggested Improvement: > Implement Natural Language Processing (NLP) or Semantic Search for the JSM Knowledge Base integration. The search engine should be able to identify the user's intent within a long sentence and filter out "stop words" to suggest the most relevant documentation based on context, not just keyword density.