Posted on

AWS Tests Enterprise Search Service Powered by Machine Learning

Amazon Kendra, a new service that uses natural language to search for enterprise customers, is now open for public preview.
Amazon Web Services (AWS), announced the Amazon Kendra public preview in March as part of its 2019 reInvent conference. The service can be used as a console application, or an API. It allows users to apply natural language queries to various content sources within their companies and receive “highly precise” answers.
According to Amazon Kendra’s product site, the service is compatible with all information sources and allows companies to “get rid information silos”. The public preview version includes connectors for Microsoft SharePoint Online and Java Database Connectivity. Other connectors for cloud-based services such as Box, Microsoft OneDrive and Salesforce will be available once the product is generally available (AWS hasn’t said when).
Other features include keyword searches and natural language queries support, the ability to retrieve results form unstructured data and FAQs, document rank, relevance tuning, and domain optimization. There are also a limited number of built-in connectors.
Analytics, query auto-completion and incremental learning are some of the new features that will be available upon general availability.
AWS will offer Kendra in two flavors: a Developer Edition for testing purposes, and an Enterprise Edition to be used for production purposes. Here is pricing information. The public preview is currently only available in the Northern Virginia, Oregon, and Ireland regions.
AWS’ announcement about Amazon Kendra closely follows similar news from Microsoft, its cloud rival. Microsoft also presented its own efforts to create AI-driven enterprise search at the November Ignite conference. Microsoft is currently offering a preview of a new “semantic” search capability to organizations.