Harvard Business Review (HBR) suggests looking for new ways to create value when deploying generative AI. Frankly, it's almost insulting that HBR needs to state the obvious. It's akin to advising people to "breathe to live." From an investor's perspective, this kind of article is nothing more than a waste of time.
The problem lies in the abstract concept of 'new value.' The sole reason companies jump into generative AI is the scent of money. Cost reduction, increased productivity, and new market creation. There's no other reason. Yet, HBR packages it as if new philosophical values need to be created. This obscures the core issue.
The core of generative AI is data. Without data, LLM or RAG are useless. Data quality, data quantity, and how efficiently data can be utilized are key to success. Before searching for 'new value,' companies must check their data strategy. Is the data pipeline properly built? Is data security solid? Is data analysis capability sufficient? If these basic questions aren't answered, generative AI is just an expensive toy.
Especially, there are high expectations for On-Device AI, but the reality is harsh. Three obstacles must be overcome: chip design technology, power efficiency, and personal information protection. The NPU performance competition is already fierce, and battery life remains a problem. Moreover, providing AI functions while securely protecting user data is a very complex problem. Simply moving computations from the server to the phone doesn't solve the issue.
It is necessary to examine past similar technology trends. For example, during the big data craze, many companies introduced Hadoop, but very few actually used the data to derive meaningful results. The same goes for generative AI. Rather than focusing on technology adoption itself, focus on three key elements: linking with the business model, data strategy, and talent development.
Looking at the competitive landscape, giants like Amazon, Google, and Microsoft are already investing vast resources to build generative AI platforms. Rather than directly competing with them, SMEs should adopt a strategy of developing solutions specialized in specific fields or utilizing the platforms of large corporations. For example, Agentic AI solutions specialized in professional fields such as law, medicine, and finance are sufficiently competitive.
In conclusion, HBR's article is misleading companies with the vague concept of 'new value.' Investors should not be misled by these pie-in-the-sky stories, but should focus on three key elements: data, technology, and business model. Generative AI certainly has disruptive potential, but if not used properly, it will only become a money pit. Let's look at the facts. The numbers will tell you where the real money is.