The automotive industry faces major challenges due to shifting market demands and technological advances, particularly in balancing innovation, cost efficiency, and growing product variability. Flexible manufacturing techniques (FMTs) with automation support adaptability to diverse component designs. Despite progress in 3D modeling, 2D technical drawings remain essential due to their compatibility with existing workflows. Automating their processing is therefore crucial for their seamless integration into modern manufacturing systems. This paper presents an AI-based approach for analyzing 2D technical drawings by detecting geometric features (e.g., reinforcements, embossments) and extracting manufacturing-relevant text (e.g., dimensions, tolerances). The prototype uses Faster R-CNN for object detection and a Keras-OCR-based pipeline for text recognition. A dataset of 204 labeled drawings covering eight feature types was created. Results confirm the system’s ability to identify complex patterns and extract critical data for downstream automation. The approach provides a scalable solution, especially for legacy components lacking CAD data and relying on scanned 2D documents.