Extracting semantic meaning from embedded visual content with Amazon Q Business
When Amazon Q Business processes your input files from a data source, it uses advanced image understanding capabilities to extract semantic information and insights from images and other visuals. For example, it can extract insights from bar charts, technical illustrations, or hierarchical organizational structures.
By extracting semantic meaning from embedded images, Amazon Q Business makes the visual information in your data sources queryable. This makes relevant information easier to find, even if it's conveyed in embedded diagrams or charts. This data provides additional context and nuance to the information in your data sources and builds a more complete knowledge base from your enterprise data.
You can enable content extraction when you add or update a data connector, or when you import a file directly. You can enable it by using the Amazon Q Business console or by using API operations. Processing documents that contain images and visuals takes more time than processing text-only documents.
When you enable content extraction for a connector, we recommend that you use incremental sync mode. If you enable full-sync for your connector, or if you import the same files individually, Amazon Q Business processes the file and corresponding images again. For more information about sync modes in Amazon Q Business, see Sync mode.
Extracting semantic meaning from embedded images and visuals incurs additional costs. For
more information, see Amazon Q Business
pricing.
Topics
End user experience
After Amazon Q Business extracts semantic meaning from images and visuals, your end users can ask questions and get answers related to the images. When an end user asks a question, Amazon Q Business retrieves relevant answers from the text and the images. Answers include the images and links for the documents that contain them.
For example, your user might ask, "Can you walk me through the process of enrolling in a health plan on the employee benefits site?" Amazon Q Business would then analyze the company's internal knowledge base, including detailed guides, process documentation, and screen shots showing the step-by-step enrollment process. It would synthesize this information into a clear, easy-to-follow response that outlined each stage of the enrollment journey.
Guidelines and requirements
The following are guidelines and requirements for extracting content from images:
-
Documents can have up to 3,000 pages.
-
The IAM service role that grants Amazon Q Business permission to access your resources must have
qbusiness:GetMedia
permissions. For a policy example, see IAM role for an Amazon Q Business web experience using IAM Identity Center -
You can use the following connectors:
-
Amazon Q Business can extract meaning from images embedded in the following file types:
-
PDF
-
PowerPoint
-
Microsoft Word (DOCX)
-
Google Slides
-
Google Docs
-