According to Andrew Ng, Agentic Document Extraction now supports field extraction, enabling users to extract specific fields such as vendor name, item list, and prices from images or PDFs of invoices and structured documents (source: Andrew Ng, Twitter, July 10, 2025). This upgrade allows businesses to automate data capture from common business documents, significantly reducing manual entry and boosting operational efficiency. The feature is expected to streamline AI-powered document processing workflows, opening new opportunities for enterprises in finance, logistics, and administration to leverage AI for automated data extraction and reduce costs.

According to Google DeepMind, the UK government has launched Extract, an AI-powered system built on the Gemini foundational model, designed to help council planners make faster decisions. Extract leverages multimodal reasoning to process and digitize complex planning documents, including handwritten notes and blurry maps, converting them into usable digital data in just 40 seconds (source: @GoogleDeepMind, June 9, 2025). This practical application demonstrates how advanced AI can streamline document processing in the public sector, offering significant efficiency gains and paving the way for further automation opportunities in government operations.