Digital Presso, an AI-based on-site data automation solution company, has established ‘AI Presso Lab,’ an AI research platform leveraging diverse conversational data generated in industrial environments, and is accelerating the development of domain-specific AI models.

AI Presso Lab is a platform designed to research and validate advanced AI technologies—such as Vision AI, LLMs, and STT—based on unstructured data including voice, video, and images generated in industrial settings. In particular, it focuses on experimenting with real-world field data and advancing models in forms optimized for industrial environments.

Through PoC (Proof of Concept) projects and collaborations with telecommunications companies and construction firms, Digital Presso is securing large volumes of call (STT) data across various industries. Based on this data, the company generates domain-specific synthetic datasets to enhance not only model performance but also real-world applicability.

AI models developed through this research aim to go beyond simple data analysis by automating an end-to-end workflow—from generating checklists based solely on call records to image classification and validation, report generation, and even payment integration.

A Digital Presso representative stated, “AI Presso Lab is not just about technical research, but about creating AI models that can be practically used in industrial environments,” adding, “Through continuous learning and validation based on real-world data, we aim to advance AI technologies applicable across various industries, including construction and manufacturing.”

Currently, AI Presso Lab is in the internal research and validation stage, with plans to expand its application scope to a wider range of industrial sites in the future.