IBM is turning towards infrastructure for AI

Silently but consistently, IBM has rebranded its Watson cognitive computing offering as AI (artificial intelligence), in line with the trend. All other major IT vendors are talking about AI, so IBM is now, too. But in the end, we must say that the promises the company had made and expectations it had provoked in this space on the customer side have simply proved too ambitious. When looking at the rough categorization into “Narrow AI” (NAI), “General AI”(GAI) and “Superficial AI” (SAI), almost all practical use cases as of today fall under the NAI category – if at all (some are just smart applications of business intelligence frameworks). As a reminder: Narrow AI refers to a conglomerate of highly diverse algorithms and techniques (e.g. neural networks and machine learning) in use today. The term GAI is used for AI capabilities that differ only marginally from human intelligence – a goal that has not yet been achieved. Consequently, the term SAI is science fiction and refers to machine capabilities that go beyond the human intellect.

Today, enterprises are only just leaving the experimental phase of AI, where only narrowly scoped projects have been tested, and start to apply AI in more general and broader use cases. These need to be integrated into companies' existing IT infrastructure. IBM is well-positioned to deliver this service and also a reference architecture, which can be used to build a home for point AI solutions from various vendors.

However, IBM is making its move in response to its fears that – despite being a clear pioneer in the field of AI – it could lose market share to others that provide more appealing and easier-to-consume AI offerings. These offerings are mostly point solutions or embedded in classic business applications, whereas the Watson offering (which aimed to be understood as general AI) was more complex and more difficult to handle.

Bottom line: Clearly, IBM is well-positioned to play the role of an AI integrator and framework company. However, it needs to sharpen its message in the AI field and create something that is easier to grasp by the market than the notion that "Watson can do it all".