Summary of Key Points
This article highlights the discussions at a gathering of AI professionals, where the focus shifted from "how to use AI technology" (such as Agents, RAG, Workflows, etc.) to "how to integrate AI into business operations and generate revenue." Five consensus points were reached:
1. Reduced technical barriers and AI as a commodity: The barrier to using AI has lowered, with models becoming more like raw materials. Fine-tuning models is less important, and selling products is harder than developing them. Behind each consensus lies a deeper insight: AI eliminates the difficulty of creating something; what truly remains as a barrier are skills in judgment, understanding users, managing knowledge, driving growth, and proving value.
2. Developing products is no longer the challenge; knowing what to develop is: Creating products used to involve multiple steps (design, development, testing), but now AI tools (like vibe coding, Coding Agents) enable anyone to quickly launch apps or prototypes. The ability to conceptualize and execute is more valuable than technical expertise.
3. AI as a component, not a product in itself: Models are no longer the end product; what matters is how they are combined with other technologies to solve problems. For example, content creation involves automating entire processes (from writing to summarizing, adding images, and creating videos). In business scenarios, this means using AI for natural language processing, data analysis, and task automation.
4. The difficulty in industry-specific AI lies in organizing industry knowledge: Instead of fine-tuning models for specific industries, a combination of general models, RAG (Retrieval-Augmented Generation), and industry-specific knowledge bases is more effective. The real challenge is managing this knowledge (defining terminology, finding authoritative sources, updating data), which is less glamorous but crucial.
5. Selling AI products has become harder: Although AI reduces development costs, attracting and retaining users remains a challenge. The focus should shift from coding to skills in identifying market needs, driving growth, building private customer bases, and setting prices—these have become essential professional capabilities.
6. Companies buy results, not just AI: Traditional industries (like steel and chemicals) care about whether AI can increase profits, save time, or reduce labor costs. Integrating AI into services (e.g., strategic consulting) provides real value, making it a profitable solution.
In summary, as technical barriers disappear, non-technical skills (judgment, user understanding, knowledge management, growth strategies, and value validation) become the core competencies for success in the AI industry. The ability to not only develop products but also ensure their effectiveness and profitability will be decisive in the future.
(End of article)
(Source: DARE to B2B)