万联易达杜新凯：AI要实现从“解答问题”到“解决问题”的跨越

Core Insights - The year 2023 is identified as the year of proactive intelligent agents, moving beyond reactive AI, emphasizing the dual engagement of industry and AI [1] - The challenges faced in the implementation of "AI + industry" include disconnection between technology and industry chains, data bottlenecks, mismatched supply and demand capabilities, and difficulties in creating a commercial closed loop [1] Group 1 - The general large models have language understanding capabilities but struggle to address specific industry problems deeply, while vertical models are limited by industry boundaries and cannot integrate the entire industry chain know-how [1] - Companies require a comprehensive AI productivity tool that can meet the needs of intelligent transformation in a one-stop manner [1] - Wanlian Yida has launched the "Wanlian Moore" AI model, which aims to cover 97 major categories of the national economy and has achieved an industry Q&A accuracy rate of over 90% through the cleaning and training of over 10 billion industry data [1] Group 2 - The development path of AI empowering industries shows a clear evolution from general (discovering potential) to specialized (solving problems) and then to new general (ecosystem empowerment) [2] - The alternating development of general and vertical large models will accelerate the formation of a new general ecosystem [2] - The proactive task orchestration and tool collaboration capabilities of intelligent agents will enhance the adaptability of large models to specific industry scenarios, achieving a transition from "answering questions" to "solving problems" [2]