Recently, smart manufacturing has been singled out as the most promising avenue in the generative AI market, especially for manufacturing powerhouses like Taiwan. However, while the concept holds great appeal, the reality reveals significant challenges in execution.
Industry surveys indicate that in the current manufacturing landscape, talent shortages and cost considerations are critical concerns, with the effectiveness of AI implementation remaining difficult to measure.
Smart manufacturing: an iceberg of obstacles
Supply chain experts point out that every investment needs to be justified, so the primary concern for most companies when adopting smart manufacturing is the potential benefits and contributions to revenue. Unfortunately, these aspects are often the hardest to quantify.
In recent years, as global supply chains undergo reorganization, smart manufacturing has evolved from a focus on enhancing in-house production efficiency and quality to addressing broader issues like cross-supply chain, cross-plant, and even cross-border operational management.
Industry insiders admit that from an operational management perspective, it's challenging to quantify management metrics, and they can usually only be measured indirectly. For example, a manufacturer might save production costs by implementing AI solutions or machine networking in production equipment, but the tangible benefits might not be directly evident at the equipment level.
Similarly, even if a company secures more orders, the benefits might not be easily quantified through metrics like "utilization rate."
According to a 2024 survey conducted by a research firm involving 150 senior managers across more than ten global industries, only 10% of companies reported real benefits from implementing AI. Meanwhile, 40% hesitated due to a lack of resources, and 50% tried but failed to achieve results.
The consensus in the industry is that developing a comprehensive method for assessing overall benefits could significantly impact companies' willingness to adopt AI. The difficulty in quantifying benefits is closely tied to the challenges companies face when implementing AI. The general agreement is that AI, when limited to executing single commands, struggles to deliver overall value.
AI effect still requires guidance; is yet to be omnipotent
According to information service providers, companies are complex organisms with interconnected components, making coordinated operations crucial. If AI can only complete isolated tasks, its value is diminished, and the benefits are hard to quantify from a single process.
Industry experts suggest that to truly harness AI's value, it must be goal-driven, completing tasks autonomously to achieve its objectives. However, AI, as it currently stands, is unable to close its loop through self-collaboration.
MIC research has found the shortage of skilled professionals to be the most challenging bottleneck faced by companies that have implemented or plan to implement AI. On the technical front, these firms also face issues of equipment integration and data integration. This underscores why AI has yet to advance from executing single commands to handling more complex tasks.
