Unlocking Business Value using Innovative SAS Solutions with Generative AI and Computer Vision

Generative AI (GenAI) has captured the imagination of executives in unlocking new business values through natural language interaction with AI using Digital AI Assistants like ChatGPT. GenAI excels at generating text and interacting with users to understand their queries. Unfortunately, GenAI has trouble answering analytical questions with precision and is prone to inventing answers or giving nonsense responses. This lack of accuracy makes it tough for companies to act on analytical results created using Generative AI. In our presentation we will be showcasing innovative solutions to solve this problem by integrating Trustworthy AI of SAS with Generative AI. In the GenAI-SAS Integration framework, all analytics results are generated by SAS in the backend, while Generative AI is used to interact with users to gather inputs to trigger the analytical models and surface results. This significantly reduces the problem of GenAI hallucinations related to analytical questions and enables companies to confidently make decisions with speed and precision.
In this session, we explore three pioneering use cases that integrate SAS Viya with Generative AI (GenAI), revolutionizing industry practices through natural language input to perform advanced analytics. First, we delve into a recipe optimization chatbot designed for Health and Life Sciences applications, utilizing a combination of open-source tools, SAS analytics and GPT-4 to allow users to input natural language commands, run a batch recipe optimization model, and explore various scenarios to find the optimal product mixture. Next, we showcase this infrastructure via another chatbot which helps non-technical users with warehouse space optimization. The chatbot interface enables users to optimize inventory movements, allocate inbound shipments efficiently, run dynamic what-if scenarios, and query results, with output data and scenario comparison displayed in an interactive SAS Visual Analytics dashboard. Finally, we present SAS DocuVision, a medical text extraction and bookmarking project that uses OCR, SAS Viya, and GenAI to create concise, high-level summaries of medical documents, highlighting essential information for quick access while reading PDFs.
To complement the GenAI-SAS integration, we will also highlight some computer vision use cases at SAS, demonstrating how these technologies can be applied to solve complex image processing challenges using SAS. These examples represent just a sample of the innovative offerings SAS has to offer, showcasing the potential for integrating advanced AI technologies with robust analytics to solve real-world problems efficiently and effectively.