At ignite 2018, Microsoft announced the Microsoft search service for enterprises for the first time, aiming to provide enhanced search experience for Microsoft teams, yammer, SharePoint, onedrive, office, windows, Bing and other products.At this year's Developer Conference (build 2019), the company announced that Microsoft search is fully available. But now it has a new competitor.
It is reported that Microsoft search, based on the Microsoft graph and artificial intelligence (AI) technology, can display the content index of organizations stored in Microsoft 365 and salesforce, Google drive, sap, servicenow, AWS and other services.
Whether you are searching for answers to people, files, organizational structures, sites, or FAQs, you can use the Microsoft search service within your organization to get answers.
Yesterday, Amazon announced its Kendra search service for businesses, aiming to compete with Microsoft search. Like competing products, Kendra supports natural language queries that help users find information across a wide range of company wide content.
In addition, Kendra also supports connectors that allow users to index what is available in services such as SharePoint online, Amazon S3, and databases.
Next year, Amazon will also provide connectors that support box, Dropbox, salesforce, onedrive and other popular data sources.
Amazon said Kendra's highlight is its ability to actively retrain and build deep learning models for enterprise datasets and employee usage patterns to improve search accuracy.
Kendra is able to fine tune its results as it is ultimately used to interact more with search results. This means that when you give a thumbs up or down on a result, Kendra will learn which results are more relevant and will prioritize them later.
In addition, Kendra provides the option to manually adjust the correlation, such as users can enhance some fields in the index (such as document freshness, View Count or specific data source) or recently accessed documents with high frequency (such as trend news and updates).
Of course, using natural language to search is not just a simple way to get the answers through keywords. Kendra will connect the points in the file to return the answer.
Whether it's text snippets, quick Q & A, or documents, Kendra can provide suggested answers in advance, rather than filtering through a lengthy list of documents to find specific answers.