Where Did the Return Go?: Why Enterprise AI Investment Doesn't Always Become Business Value — and What to Do About It

Where Did the Return Go?

Why Enterprise AI Investment Doesn't Always Become Business Value — and What to Do About It

Your AI dashboard is green. Your teams are faster. Your output is up.

So where did the return go?

Enterprise AI rarely fails because the technology does nothing. The harder problem is that organizations often measure the wrong thing. A faster task is not automatically a better process. More output is not automatically more value.

Where Did the Return Go? follows Maya, an experienced enterprise leader trying to understand why an apparently successful AI investment is not producing the business value everyone expected.

The question becomes:

What happens between AI capability and actual enterprise value?

AI enters an existing organization—with processes, handoffs, approvals, legacy technology, incentives, governance, human judgment and verification requirements.

The book moves beyond:

“How much faster did AI make us?”

and asks:

“How much of that improvement actually became business value?”

Maya discovers that productivity gains can disappear inside organizational friction. A team can complete tasks faster while the overall process becomes more complicated. More AI-generated output can create more verification work. A workflow that appears cheaper can introduce new costs for integration, governance, security, monitoring and human oversight.

And there is another cost that is easy to overlook: the cost of AI itself.

Token and inference consumption, model usage, retrieval, tool execution, infrastructure and model management can materially change the economics of an AI investment. These costs belong in the ROI calculation.

Another question follows:

What does the organization now depend on to keep producing the return?

External models and platforms introduce dependencies involving availability, pricing, model behaviour, APIs, policies, capacity and provider roadmaps.

Drawing on research from economics, organizational science, productivity, human factors and generative-AI research, this book explores the gap between technology investment and business value.

It examines why AI productivity does not automatically become ROI; why coordination can absorb efficiency gains; why verification becomes critical; why human capability remains essential; why complementary investment matters; why task-level productivity can mislead; how AI operating costs affect returns; why external dependencies belong in investment decisions; and how leaders can distinguish AI output from accepted business outcomes.

At the heart of the book is a simple proposition:

AI ROI is not the value produced by the model. It is the value that remains after the organization has paid for, coordinated, verified, governed, operated and depended on the system required to produce it.

This is not a book arguing that AI is overhyped. It is about what happens between capability and adoption, productivity and value, output and outcomes, investment and return.

For CIOs, CTOs, CFOs, transformation leaders, enterprise architects and AI leaders, Where Did the Return Go? offers a different way to think about enterprise AI.

Because the question is no longer simply:

“What can AI do?”

It is:

“What has to remain true for the value AI creates to actually reach the business?”