OCR has been the pinnacle of modern technology for the last twenty years, but is the problem of converting images to text solved? Is the field of optical character recognition dead? In this presentation, we will look at the current state of the art and how bright the future of OCR is.

Didn't get the response you wanted from ChatGPT? Maybe you are the problem! In this session, we will introduce the topic of Prompt Engineering and look into ways scientists and industry try to learn how we should format prompts and ask questions for LLMs. We also cover LLM agents, where prompt engineering and well-written system prompts are essential for the successful integration of LLMs with your external systems.

This session will cover modern GenAI and traditional AI service approaches in solving document management cases, such as classification and information extraction for invoices and legal documents. We will cover the importance of data confidence provided by more traditional approaches versus generative data, which may need manual review. Furthermore, we will work around the limits for both approaches and discuss cost efficiency.