The Real Reason AI Isn’t Delivering Business Value

Dhiraj Rajaram, is the Founder & CEO of Mu Sigma and a tech visionary. Mu Sigma is one of India’s early unicorns.

Every day, it seems there is a new study highlighting that AI deployments haven’t paid off for businesses. None of this should be surprising.

​AI is great. The technology is impressive and gets better every day. The problem with AI, however, is not the technology itself. Rather, it’s the unaddressed organizational deficiencies that limit an organization’s ability to take real advantage of it. Unfortunately, these are much bigger, more structural issues that need to be addressed. Organizations are good at capturing what happened but much weaker at preserving how reality was perceived, why decisions were made and what was learned. AI is exposing that gap, not creating it.

​Here’s what I mean.

​Spaghettification

​A core problem limiting AI’s value in an organization is the unaddressed complexity that any large organization experiences. As a Fortune 500 company becomes more and more successful, it adopts layers and layers of unaddressed complexity. This creates what I call the “spaghettification” of the organization.

​Take, for example, a large footwear brand that has expanded into new channels or geographies. It will have more transactional data as a result, which is good. But insights into buying patterns become increasingly fragmented, making measurement and progress tracking difficult. Of course, this spaghettification came about for valid reasons. It came from success, and great companies are the most spaghettified. But it also creates complexity at the data layer, the IT layer and, most importantly, the insights layer.

​It’s very hard for new technology to land in a spaghettified world. Infrastructure, standards and better data tagging matter, but they only go so far. Even with clean data, organizations can lose the assumptions, interactions and judgment behind decisions. The problem is not just fragmented data; it’s also fragmented context.