The Asset Administration Shell (AAS) is a cornerstone of Industry 4.0 and the Digital Product Passport, offering standardized digital representations of industrial assets. Although manufacturers already maintain extensive technical product documentation, creating AAS instances from heterogeneous PDF datasheets remains labor‑intensive due to varied document structures and company‑specific terminology.\ \ We introduce AAS‑RAIL, a retrieval‑augmented information extraction approach that automatically generates AAS from PDF datasheets using large language models (LLMs). Instead of relying on a fixed set of few‑shot examples, the proposed retrieval‑augmented in‑context learning (RAIL) retrieves LLM‑generated extraction helpers from similar AAS instances, providing instance‑specific context. This enables the model to adapt to company‑specific naming conventions and formatting styles without any fine‑tuning.\ \ The main contribution is the dynamic selection of company‑specific AAS examples for each datasheet, replacing static prompting with an adaptive pipeline that combines semantic retrieval and structured extraction. Evaluation on a collection of industrial product datasheets using both open‑ and closed‑weight LLMs shows that RAIL consistently improves extraction quality by 30.4%–52.4% over conventional few‑shot prompting, demonstrating its effectiveness for company‑specific AAS generation.\ \ Review: The approach markedly boosts the accuracy and robustness of AAS generation for individual companies, offering a practical solution for automated industrial document processing.