Did you miss the Ask the Expert session on How Can I Process Scanned Document Images in SAS Viya? Not to worry, you can catch it on-demand at your leisure.    Watch the webinar   This session will provide you with an introduction to Document Vision capabilities, how SAS orchestrates document processing and analyzes document images. You will also learn about real-world examples of where and how Document Vision can be applied.   Topics Include: Document processing, text analytics, and the challenges involved. How SAS can effectively orchestrate and enhance complex document vision processes. Use cases and examples of where SAS's document vision process can be applied.   Below are some highlighted questions from the Q&A segment held at the end of the session for ease of reference. I’ve attached the slides as well.    Is SAS document analysis a separate license from VDMML or VA? Or included in those licenses?  SAS Document Analysis is offered as a separate add-on license to read in document images and produce analysis ready data. VA, VDMML or VTA will be a separate license, depending on the use case, to analyze this data. SAS can work with customers to determine any additional software or services that may be required.    Can this translate the Hindi language pdf to English language document?  If the OCR engine that you are using can read and transcribe Hindi, SAS can analyze Hindi documents in its native language. SAS provides out of the box NLP functionality in 33 different languages to enable native language analysis, including Hindi. SAS has a dedicated global linguistics team that produces language specific dictionaries and assets. Since the grammar rules across languages differ, this approach is much more reliable and accurate than translating non-English feedback into English and then running them through a text analytics engine that only supports English grammar rules. If you would still like to translate the text first, that will need to happen outside of SAS.    Where and or in which SAS module does the OCR tool live?  SAS recently released SAS Document Analysis as a tool on top of SAS Viya, which orchestrates the OCR process. Currently, the OCR technology lives in the customer’s environment, and SAS Document Analysis can call the OCR endpoint either using the API or local container for additional data security.    Is Document Vision independent of say SAS Model Studio and/or SAS Studio?  Document Vision leverages SAS Document Analysis, SAS Viya, and professional services for any customizations needed. SAS Studio and Model Studio are components of that process. SAS Studio flows are leveraged to stitch together pieces of the process and schedule batch execution. Model Studio is used for advanced data and text analytics.    Is OCR a process that can be done internally? Or it’s only really feasible by reaching out to y’all? SAS requires access to a third-party OCR engine that is customer licensed. While access to other OCR engines can be configured, SAS Document Analysis currently offers built-in support for Microsoft CV and AWS Textract OCR service endpoints.  SAS hosted OCR capability is on the roadmap. While you can process the OCR internally, SAS performs additional pre-processing steps like page rotation, converting file types, and resizing the source document to optimize for the OCR endpoint. This ensures enhanced accuracy in OCR output and document extraction.     Is this software free with SAS Studio? Or does it have a cost?  Document Vision is an eco-system of SAS products and services, which depend on the individual use case. SAS Document Analysis is an add-on software product and a pre-requisite for Document Vision if you are dealing with scanned document images. SAS Document Analysis is not free and does not come with SAS Studio.     What exactly is SAS doing in these apps from the tool vs how much coding is happening in the background?  This depends on the particular use case. Document Vision has different components and capabilities that can be customized as necessary. SAS Document Analysis is a tool that provides out of the box functionality to read scanned document images and produce analysis ready data. For this, users can choose to code or use SAS Studio custom steps. This data can then be consumed in SAS Viya for analysis, text extraction, document classification, etc. For more enhanced and customized workflows, coding may be required.    Is Document Vision the same thing as Visual Text Analytics?  No. Document Vision is not the same thing as Visual Text Analytics. Visual Text Analytics is a software component in SAS Viya that allows users to leverage natural language processing (NLP) and other capabilities, such as topic detection, document categorization, concept extraction, etc., to analyze machine readable text. Document Vision leverages Visual Text Analytics among other SAS software and services, as needed, to read and analyze scanned document images that are not inherently machine readable.      Can we use SAS Viya for text analysis (e.g., content analysis) after completing OCR process or using the original images (unstructured ones)?  You can use SAS Visual Text Analytics or other text capabilities in SAS Viya after completing OCR on document images and transcribing those images to machine readable text, but you cannot use the original document images directly. SAS recommends using SAS Document Analysis for converting the document images into machine readable text and generating analysis ready data to ensure enhanced accuracy in OCR output and analysis results.     Can you highlight a bit on the Microsoft | Document Vision | AWS stacking?  Document Vision is an eco-system of capabilities that leverages SAS software, SAS services where needed, and third-party OCR. This is where Microsoft and AWS come into play. SAS leverages Microsoft and AWS OCR capabilities out of the box, while primarily focusing on providing data ingestion, processing, analytics, and workflow automation. While other OCR engines can be accommodated and are on the roadmap for out-of-the-box support, Microsoft and AWS OCR provide flexibility and accuracy with handwritten documents and noisy document images. They have shown excellent performance relevancy. SAS post-processes this OCR output, generates additional metadata and metrics, and performs intelligent document extraction.    Do you leverage any Generative AI / LLM in your solution?  The use of Gen AI Large Language Models for Document Vision is under research. SAS is investigating the feasibility and performance of doing so. Large language models can be very helpful, but they can also be very expensive. SAS wants to ensure that we use them in an intelligent, cost effective, scalable, and reliable manner, and that they are producing reliable and trustworthy results. Some of the areas where we're exploring the use of large language models with Document Vision are synthetic data generation, conversational user interfaces, data extraction, fuzzy matching, and automating e-mail communication.    Recommended Resources: Document Intelligence: The Next Big Thing In AI SAS AI for Medical Record Review:  Streamlining Medical Record Review with SAS (includes demo) A Digital Assistant for Medical Record Review SAS Viya: SAS Viya Overview    Want more tips? Be sure to subscribe to the Ask the Expert board to receive follow up Q/A, slides and recordings from other SAS Ask the Expert webinars. Just hit SUBSCRIBE here:       ... View more