Amazon Kendra is the right choice for you if your Alfresco repository contains documents that you want to ask questions of, like knowledge bases, bookstores, or archive catalogues.
Amazon Kendra is an intelligent search service powered by Machine Learning. The service is fully managed, so there are no servers to provision, and no Machine Learning models to build, train, or deploy. Amazon Kendra enables the use of natural language questions to get an exact answer from within a document, whether it is a text snippet, FAQ, or PDF document. In addition to keyword searching, Amazon Kendra can answer questions expressed in natural language, like "How does Alfresco Transform Engine work?".
Amazon Kendra uses machine learning to improve search results over time based on search trends and feedback from end users. Amazon Kendra learns from user interactions and input to promote favourite articles to the top of the list to find the best relevant document for that topic.
Since Amazon is providing an Alfresco Connector for Amazon Kendra, a set of selected documents from Alfresco can be indexed by Kendra to support this kind of questions. Read permissions on the content are indexed in Kendra as well, so the results are filtered according to user authentication.
The following diagram describes an Alfresco deployment connected to the managed service Amazon Kendra.
Alfresco can be used to find documents using keywords or technical search syntax like Alfresco Full Text Search (FTS) and Content Management Interoperability Services (CMIS). While Amazon Kendra can be used when using natural language questions. The combination of both services helps to complete searching capabilities when deploying a Content Service Platform solution.
Additional details on the configuration and use of this architecture are available in AWS official documentation:
Happy natural searching!
Would you like to know more? Would you like to see a deeper integration between Alfresco and Kendra? Let us know in the comments below!
