- Andrew Ng states that Agentic AI is a major step forward as AI no longer just generates content but also executes complex sequences of actions.
- Unlike Generative AI, which acts as a “copilot,” Agentic AI can automate multiple steps in a workflow and operate parts of a business.
- The greatest value comes not from a 3–5% efficiency improvement, but from redesigning entire processes for massive growth. For example, instead of shortening a minor step, businesses can process records in 10 minutes instead of 1 week, creating a new competitive advantage.
- The biggest challenge is reliability: systems can be built quickly in a few weeks but require months to reach enterprise standards.
- Data does not need to be perfect from the start; it should be improved gradually based on each practical application.
- AI can automate about 30–40% of tasks, but humans are still needed for the remaining 60–70%.
- People who know how to use AI will replace those who do not, creating significant pressure on training and skill upgrading.
- The cost of building Agentic AI is decreasing as AI models become more powerful, making deployment easier over time.
- Business leaders need to understand the technology and lead AI strategy from the top down to create real value.
📌 Conclusion: Agentic AI marks a shift from support to autonomy, capable of automating 30–40% of work and restructuring entire business processes. According to Andrew Ng, the great value lies not in small improvements but in breakthrough growth, such as reducing processing time from 1 week to 10 minutes. However, challenges regarding reliability, data, and personnel skills remain significant, requiring long-term investment and a clear strategy from businesses.
