Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{192164,
        author = {Anuj Singh},
        title = {Enhancing OCR in Power Automate Using Preprocessing and Large Language Models (LLMs)},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {9},
        pages = {295-301},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=192164},
        abstract = {Optical Character Recognition (OCR) plays a crucial role in streamlining document processing workflows. Power Automate, which is Microsoft's platform for automating tasks, integrates OCR features through AI Builder and various connectors. That said, the accuracy of OCR can be impacted by factors like image quality, complex layouts, and unusual fonts. This paper delves into how we can enhance OCR accuracy and improve text interpretation in Power Automate workflows by integrating preprocessing techniques and Large Language Models (LLMs). 
We suggest a hybrid method that combines image preprocessing, structured data extraction, and contextual validation with LLMs, resulting in notable gains in text recognition accuracy and the reliability of automation. In the preprocessing phase, we utilize OpenCV techniques such as denoising, skew correction, and adaptive thresholding to enhance the quality of raw images before they undergo OCR. After that, the output from OCR is fed into an LLM, which refines, validates, and organizes the extracted data.},
        keywords = {Optical Character Recognition (OCR), Robotic Process Automation (RPA), Large Language Models (LLMs), Power Automate (PA), Azure Functions.},
        month = {February},
        }