Unlike the search tools that require significant capital expenditure before implementation and continuous
manual review and calibration, ML-powered intelligent search-as-a-service can be implemented quickly and
bypass adoption barriers. The key differentiations include no special customization or tuning required; the
technology can be applied out of the box to different use cases across a variety of domains; and it helps deliver
answers and actionable insights extracted from complex technical content.
Amazon Kendra is an excellent example of an on-demand intelligent search solution that delivers an effective
answer-oriented experience, powered by content unification, natural language understanding, text analytics, and
machine learning.
The intelligent search difference: engineered cloud-first, ML-first
Machine learning and artificial intelligence are making a phenomenal impact on the
search implementation and adoption process to deliver an adaptive search experience:
1. Intellectual interpretation of search query for enhanced user intent detection.
2. For general queries, accurate answers, and pointers to the most relevant passage from
unstructured document content.
3. Identify and deliver clusters and facets of related knowledge mined from across the set of
diverse, disparate, and distributed repositories to enable smart navigation and discovery of
relevant information.
4. Interpret user query patterns and behavior to deliver enriched content (e.g., auto
generation of synonyms) and corresponding search experience.
5. Deliver meaningful results to users without needing any significant content curation, query
customization, or other forms of upfront tuning.
10
