Data extraction is important to understand with your vendor as it significantly impacts time to value, scalability, and the longevity of your automation solution. It's the first step in automating invoices. How Medius captures data in the tool is unique. We do not take a traditional template based approach, which can historically be limiting. So what we do instead is that we read the invoice using AI, which is designed to pick up on both header and line item information in many different languages. Our AI recognizes relative location of the data, meaning we can find information even when the layout changes and items move around the page. You can configure fields differently for suppliers because the fields that I capture might be used on some layouts for matching and routing purposes, but maybe it's not needed on all. When I make changes simply by clicking into a field and clicking into the document, Medius remembers the changes that I make so that I'm able to automatically extract this data correctly on future invoices. The Medius needs verification for the information extracted. Maybe it's an old invoice state. For example, yellow triangles direct me to the specific fields in question, allowing me to resolve the issue, and the system continues to learn throughout my actions. Again, once you're up and running with the solution, this queue of invoices will be for exceptions only. The majority of invoices will flow directly into the workflow without stopping here. When you look at automation, we wanna make the entire process more efficient, not just replace today's tasks with data extraction. And the combination of first time identification and machine learning allows your AP team to spend more time on strategic initiatives rather than unnecessarily keying data fields.