This work investigates the unstructured documents that contain handwritten texts and normal text, which are digitized using the applications of artificial intelligence. Because of the varied handwriting quality and the unstructured nature, it is still a challenge to achieve better efficiency in digitizing unstructured and handwritten documents. To determine the best technique for converting handwritten text to a digital format, the work assesses the various Optical Character Recognition (OCR) models like the Document Text Recognition (DocTR), Easy Optical Character Recognition (EasyOCR), Python-tesseract (Pytesseract), and Keras Optical Character Recognition (KerasOCR). Also, explore the work that has been done so far related to Kannada Epigraphy. The mentioned models were tested using a suitable dataset. To check the accuracy of the model’s performance metrics, like the Character Error Rate (CER) and Word Error Rate (WER) were used. The results showcase the unique strength of each model. The Document Text Recognition (DocTR) and Keras Optical Character Recognition (KerasOCR) based on their advancement in architectures, have proved their efficiency in unstructured and complex documents. Easy Optical Character Recognition (EasyOCR) proved its robust recognition of multiple languages. The Python-tesseract (Pytesseract) prevailed with very minimum errors in converting the high-quality images to digital text. The work compares the performance of all the OCR models in managing the unstructured handwritten documents and has also explored the various works done on the Kannada Epigraphy. This provides the way for the forthcoming inventions in the field of Kannada Epigraphy document processing.