Why Do Big AI Rollouts Keep Failing?

Matt joins Paul for another episode in the “Seven Big Questions About AI” series to explore a provocative question: why do big AI rollouts keep failing? They unpack whether the widely reported failures of generative AI projects are really about the technology itself, or whether organizations are approaching AI with the wrong expectations, metrics, and assumptions about what transformation actually means.

The conversation explores why traditional measures of ROI may not fully capture the value of AI, particularly when organizations expect immediate, measurable business results from a technology that is changing how work gets done. They also examine the enormous hype surrounding AI, the fear of falling behind competitors, and the growing backlash against AI from people who see it as disruptive, unnecessary, or simply producing “slop.” Paul and Matt discuss how these reactions can obscure a more important question: what are organizations actually trying to transform?

They then compare AI transformation with the digital transformation that preceded it. While digital transformation was largely a technology transformation with a human component, Matt argues that AI transformation fundamentally reverses that relationship. AI is a human transformation with a technology component, because it has the potential to change what people do, how they work, how teams operate, and ultimately how organizations are designed. The conversation explores why simply adding AI training or asking employees to adopt new tools is unlikely to be enough, particularly when many people never wanted the technology in the first place.

Paul and Matt also discuss the risks of approaching AI transformation through the same playbook used for digital transformation, including the tendency to treat people as an afterthought and the temptation to reduce headcount before understanding how AI will actually change the organization. They explore why organizations may be better served by slowing down, understanding the value of the capabilities they already have, and resisting the constant pressure to deploy the newest AI technology simply because competitors are doing so.

The conversation ultimately turns to what organizations may need to look like in an AI-driven environment. Matt argues that traditional organizational structures may need to become more flexible and adaptable, with teams forming and disbanding around specific problems and opportunities rather than remaining fixed indefinitely. Using the movie industry and Christopher Nolan’s “The Odyssey” as an example of this kind of “swarming,” they explore how AI could enable organizations to assemble focused teams around ideas, create value quickly, and then reorganize for the next challenge. The episode ultimately argues that big AI rollouts fail not because organizations lack access to powerful technology, but because they are approaching a fundamentally human transformation as if it were simply another technology implementation.

Resources Mentioned:

Understanding Organizations...Finally! — Henry Mintzberg
Amazon: Get the book here!

Designing for Change: Time for a Future of Work Executive
Article: Read here!

Humanity Working is a podcast focused on helping individuals, teams and organizations be ready for the future of work by maximizing their human potential.

For more information, and access to our weekly newsletter, visit us at humanityworking.net.