Question · Q2 2025
Zhang Huang asked for quantifiable metrics or numbers regarding AI-driven R&D efficiency improvements, such as development cycle or per capita output. He also inquired about the future optimization potential for the R&D expense ratio and the company's next focus areas for efficiency enhancement.
Answer
Chairman and CEO Fu Sheng stated that specific detailed R&D efficiency data is considered core competitive intelligence and cannot be fully disclosed. However, he provided an example: a recording summary tool that previously required a 5-10 person team was developed by two less-experienced colleagues using AI, demonstrating significant efficiency gains. He linked the company's rapid reduction in losses (from RMB 200M+ to near break-even after acquiring OrionStar) to both business growth and R&D efficiency. He clarified that optimization isn't just about cutting R&D expenses but improving team efficiency to support better products and revenue. The primary challenge and next focus for efficiency is building an 'AI-native' organization, where employees deeply understand and utilize AI, aiming to reduce role specialization and enhance individual capabilities across product management, development, and various engineering roles.