Everything Auto-Complete

The emergence of large-scale AI, particularly ChatGPT, has sparked significant interest in artificial intelligence. However, this journey is just beginning. Many productivity innovations have been discovered and shared through ChatGPT, focusing on workflows that gather and process information. Despite this, not everyone has seamlessly integrated these workflows into their daily lives, making the democratization of AI an ongoing process.

Looking at areas where AI is already widely democratized, we can glimpse the future of AI-driven revolutions accelerated by generative AI, centered around 'auto-completion.' Features like search query auto-completion, email auto-completion, and text message auto-completion are widely used because everyone uses emails and texts. This workflow involves text input and output, with auto-completion understanding and predicting text. This represents online auto-completion. The AI revolution is expected to intensify when offline auto-completion becomes more concrete and tangible.

Currently, there are various expectations for the robotics industry. Ultimately, if a robot can understand situations through sensors, utilize a powerful reasoning engine like large-scale AI, and operate specific workflows, it fits the definition of a robot. Therefore, all manufacturing industries that create products interacting with people will fall under the scope of the robotics industry. The realization of these expectations may not necessarily involve human-like robots.

This concept can be applied universally because AI can process unstructured data beyond text and understand context. Furthermore, its ability to assess situations and reason based on given inputs will become faster and stronger.

How should companies prepare for this change? The answer is 'auto-completion.' For example, consider ordering coffee at Starbucks. Currently, users need to go through several steps to recharge the app. In the future, the Starbucks app could recognize that the user has a zero balance when opening the app at the counter and automatically suggest a recharge. The coffee machine could also recognize the order and prepare the coffee automatically. The same applies to the kitchen. When a bag full of groceries arrives, AI can recognize items that need refrigeration and automatically open the fridge door. A box for recycling plastic bags could approach and pick them up.

Currently, this kind of automation is possible with rule-based systems, but understanding various situations comprehensively and issuing precise commands to each machine remains challenging. However, the technical bottleneck hindering this has been resolved, which is the essence of current large-scale AI and multi-modal systems.

Companies need to create valuable 'auto-completion' customer experiences. To do this, they must understand the lives of their customers and the workflows surrounding their products. It is unnecessary for manufacturers to develop reasoning engines themselves. Instead, they should focus on how to provide information to large-scale AI, which will understand the workflows, communicate with each product/service, execute rules, and create a seamless and natural customer experience.