CLM - Modernized Search - Improve Search to Handle Case and Quote Variations in Phrase Searches

Problem / User Pain Users performing exact/phrase searches in CLM can miss relevant results when the document text differs from the query by:

• Case (e.g., searching agreement's records vs. document text AGREEMENT’s records)
• Punctuation style (e.g., straight apostrophe ' vs. curly apostrophe ’, and straight vs. curly double quotes)

This is especially confusing in quoted phrase searches, where users expect “exact phrase” matching but not failure due to common formatting differences introduced by OCR or word processors.
 

User Expectation Because CLM functions as a document repository, users reasonably expect search behavior similar to word processing tools:

• Curly and straight quotes/apostrophes should be treated as equivalent for matching.
• Exact/phrase search should support a case-sensitivity toggle, so users can choose whether case must match exactly.
   

Proposed Enhancements

1. Case-Sensitivity Toggle for Phrase Searches

   • Add an option (e.g., checkbox or advanced search setting) to enable/disable case-sensitive matching for exact/phrase searches.
   • Example: Allow "agreement's records" to match AGREEMENT’s records when case-sensitivity is off.

2. Normalize Quote/Apostrophe Variants in Search Matching

   • Treat punctuation variants as equivalent during indexing and/or query parsing:
     • ' and ’
     • " and curly double quotes (e.g., “ and ”)
   • Ensure phrase searches do not fail solely due to punctuation style differences.
      

Benefits

• More intuitive and consistent “document repository” search experience
• Fewer false negatives in search results (especially in OCR-derived text)
• Reduced support burden and improved user confidence in CLM search accuracy
   

Interim Workarounds (until implemented)

• Remove quotation marks to broaden results
• Try both curly and straight apostrophe styles when searching exact phrases
  
  
  [Internal reference: 16843931]