Features & Roadmap

Using Optical Character Recognition (OCR) and Large Language Model (LLM) technology, AISAY extracts and interprets information from a wide range of documents including handwritten notes, printed articles, multilingual text, and both structured and unstructured formats. Common documents include identification cards (NRIC/Passport), bank statements, invoices and agreements.

AISAY is able to process up to 50 documents simultaneously, reducing wait times for users.

Users can save frequently used prompt templates and share them with team members through the web application.

Users can review flagged PDF annotations to spot potentially suspicious elements, such as shape or text overlays that may conceal or amend original content, so that they do not go unnoticed when document processing gets automated.

Introduce fine-tuned vision language models (VLMs) on top of existing OCR and LLM capabilities, so that users can potentially process complex documents containing visual elements such as images, charts, and diagrams across a broader range of document formats.

Enhance prompt engineering with an intuitive web interface that simplifies prompt creation. Users can define their desired table output, and the system will automatically craft an optimised prompt, helping them achieve effective results on their first attempt with minimal effort.