Instead of worrying about driving ROI from GenAI pilots – time to face larger concerns

A recently published MIT report shows that 95% of generative AI pilots fail to create business value. Only a small minority of projects deliver ROI or noticeably transformative business impact.

The missing link is not the technology itself, but a persistent knowledge gap that creates an inability to understand how generative AI can meet the requirements and needs within the organization.

Many companies make costly investments in generic tools that cannot be adapted to internal processes. Or they attempt expensive in-house builds that perform worse than more specialized solutions from tech consultants.
This "GenAI gap" creates two camps: one with a few leading companies (including startups) that focus on concrete and actual business challenges and realizing fast revenue gains; and the other, the majority, stuck with high frustration over the lack of value.

But expecting demonstrated value at the pilot stage may lead us astray. Instead of worrying about driving ROI from GenAI pilots, companies could focus on identifying a handful of the most promising AI projects and ensuring that their organizational structures, talent, governance, and data infrastructure are fit to scale them.

In any case, productivity will come. Perhaps like a ketchup effect once companies and societal institutions figure out the tricks, which lie not in the technology itself but in humans’ trust in it and in adapted organizational structures. But the uncomfortable truth is that these likely historically massive productivity gains will not be evenly distributed. Research shows that AI drives polarization of incomes and profits by:

• Disproportionately favoring capital and large corporations with resources
• Generating greater gains in wealthy countries than in poorer ones
• Reinforcing divides between sectors and professions

So while we dream of productivity gains, let us truly address how humans can lead AI.

Not the other way around.

‍