by Max Smolaks 12 February 2020
American startup Luminoso has launched an AI-based product aimed at improving the results of search engines used by businesses.
Search Enhancement relies on the company’s proprietary natural language understanding (NLU) tech to supplement existing search functions, automatically producing better results in response to complex or poorly defined queries.
A more understanding search engine
Luminoso has been working on machine learning tools for language analysis since 2010. The company was spun out of a research laboratory at the MIT and is headquartered in Cambridge, MA.
Anyone who has dealt with search engines embedded into enterprise applications and websites will admit that that they leave a lot to be desired, when compared to user experience delivered by consumer-focused services from Google, or even Microsoft.
Search Enhancement works by expanding enterprise search engine indexes. Once configured through an API, it augments keywords into lists of related topics, and offers suitable responses to searches that contain misspellings, acronyms, domain-specific terminology, slang, and words with multiple meanings.
“Search Enhancement not only finds exact terms, but also any other related ways people might express that search term, resulting in the best search results regardless of the domain or industry,” explained Adam Carte, CEO of Luminoso.
The transfer learning-based system can be used to augment any type of search engine, including those used for e-commerce sites, intranet systems, or wikis. For example, Japanese automotive parts giant DENSO used it to create a search system across more than one million maintenance-related documents.
“While a number of AI-powered search engine offerings claim to provide more intuitive query results, none offer the ability to make an organization’s existing search system automatically understand millions more words and phrases,” Carte said.
“Search Enhancement works right out of the box, offering organizations across industries the fastest way to start getting more human-like and useful results from their existing search engine solutions.”