Coxwave , operator of the AI product analysis platform Align, announced on the 1st that it has been selected for the "2025 Scale-Up TIPS R&D Global" program, hosted by the Ministry of SMEs and Startups. With this selection, Coxwave will receive 1.5 billion won in research and development funding over the next three years.
The Scaleup Tips R&D Global program supports the scale-up and global market entry of promising small and medium-sized venture companies. Coxwave was selected in recognition of its AI monitoring and analysis capabilities, as well as its global expansion capabilities, accumulated through its operation of the AI product analysis platform "Align."
Through this project, Coxwave will develop "multi-action sequence analysis and anomaly detection technology for LLM-based agent services." While existing commercial solutions have focused on observability, tracking the point of error occurrence, Coxwave plans to build an error response system using a risk detection method that proactively identifies high-risk cases even when operating normally.
This technology simultaneously verifies the agent's final response and the generation process. The final response stage identifies high-risk factors, detects unfounded generation, and detects data distortion. During the generation process, it detects abnormal patterns such as unintended behavior, inconsistencies, and incorrect tool calls. Furthermore, it focuses on multi-turn analysis of real-world work environments, rather than single requests, and develops models capable of handling long-term contexts.
Coxwave is pursuing a strategy to significantly reduce operational resources compared to a single LLM by combining lightweight models and machine learning techniques, leveraging LLM lightweight technology acquired through collaboration with NVIDIA. The company announced its goal of leading the global AgentOps market with its latent error detection technology, addressing the growing demand for systems that enhance the operational stability of AI agents in industries requiring high reliability, such as finance, healthcare, law, and manufacturing.
“This selection is a recognition of our technological prowess in securing AI agent reliability,” said Kim Joo-won, CEO of Coxwave. “We will pioneer a new market in the field of AI reliability verification and establish global standards through a differentiated approach that detects potential errors in advance even during normal operation.”
- See more related articles
You must be logged in to post a comment.