Lab 2: Keyword Search vs Semantic Search

Same query, two search engines. See why exact keyword matching isn't enough.

🔍

Problem #2: Keyword search misses what you mean

Traditional search only matches exact words. Semantic search understands meaning. Try the guided queries below to see the difference.

The Document Corpus (15 NovaCorp policy sentences)

#1You are entitled to 21 days of paid annual leave per year at NovaCorp

#2Staff can claim up to $500 for gym memberships and fitness classes annually

#3Remote work is allowed up to 3 days per week for Level 3 and above employees

#4Mental health counseling sessions are fully covered — up to 12 sessions per year

#5Free basement parking is available for all full-time NovaCorp employees

#6Sick leave beyond 3 consecutive days requires a doctor's certificate

#7Paternity leave of 14 days is fully paid and must be taken within 3 months of birth

#8Performance reviews are conducted twice a year in June and December

#9Employees working on public holidays receive double pay or a compensatory day off

#10Wellness Days cannot be combined with annual leave and expire at end of each quarter

#11NovaCorp offers 26 weeks of fully paid maternity leave

#12Level 1 and Level 2 employees may work remotely 1 day per week after 6 months

#13EV charging stations are available on B2 level, bookable via the facilities app

#14Unused annual leave can be carried forward up to 5 days into the next year

#15Employees on probation are entitled to only 7 days of leave in the first 6 months

Try these guided queries

Takeaway: Keyword search fails when users don't use the exact same words as the document. Semantic search understands meaning — "time off" finds "annual leave" because they mean the same thing. This is why RAG uses semantic (vector) search, not keyword search.