Member-only story
AI for Enterprises
Why This Article Matters
The commercialization of AI applications, particularly in the areas of large language models and computer vision, has predictably given rise to a new generation of startups and a wave of interest from enterprises. As with previous tech trends, business buzz exceeds business understanding, and the industrial landscape will soon be littered with the corpses of failed initiatives, the craters of missed opportunities, and the smoke of illusory returns.
That is not to say that AI is not the most important technological trend of the moment. Rather it is to say that an understanding of where AI can be most effectively commercialized, while essential, is neither obvious nor intuitive. It requires insight into both AI’s current capabilities and limitations.
AI for the Business Mind
A way to think about AI from a business perspective is this: AI is a pattern recognition and repetition engine. Machine learning, which is the computing process through which AI is delivered, can be simplistically described as identifying patterns in one large data set (the training set) and iteratively applying those patterns to another large data set (the testing set) until the patterns are validated as useful.
Importantly, AI is probabilistic rather than deterministic. This means that AI can produce unexpectedly creative results. In applications where creativity is useful, this is good news. In applications in which there is no room for error, this is bad news.
Everybody Was (Mostly) Wrong
At the dawn of AI, most everyone assumed that AI was coming for the manual jobs first: the lower wage industrial and administrative workers doing repetitive tasks. A computer would have an easier time moving boxes or filling out a spreadsheet than ideating designs or writing presentations, right?
Wrong.
There is no room for error when it comes to moving a box or filling out a spreadsheet. It has to be done the same way every time. AI’s probabilistic nature is a nightmare for such situations. On the other hand, creative endeavors that have more than one “right” answer are a perfect fit. Especially when there is a human validating the output.
