The questions nobody planned for

For years, organizations have been asking what AI can do. Can it answer the customer? Resolve the ticket? Review the contract? Complete the workflow without someone directing every step?

Increasingly, the answer is yes. That is where a different set of leadership questions begins.

• What happens to the time people get back?
• If AI handles the routine work, does the work that remains become harder?
• If one team suddenly produces twice as much, what happens to the next team in the process?
• When AI moves from recommending to acting, who decides how much authority it should have?
• Six months after the dashboard turns green, how will you know whether any of this actually made the business better?

None of those questions means the technology failed. They exist because it worked.

You may own the service desk, the support function, a shared-services team or the operations behind a business unit. Either way, you are the one who will be asked what should happen next. This book is built for that moment.

The first result

The first result may be right without being complete

Suppose AI gives a team 20 percent of its time back. That’s real. But what should the 20 percent become? More output? Better service? Growth? Employee development? Lower cost? Fewer people? Or will demand simply consume the capacity before anyone realizes it was there?

The productivity gain cannot answer that question. Leadership has to.

Now imagine AI makes one part of a workflow dramatically faster. The dashboard looks excellent. Then another team starts struggling. Did AI fail? Not necessarily. The first improvement may have been completely real.

The consequence simply traveled.

The evidence

A real gain still leaves a decision

When researchers followed 5,172 customer support agents as a generative AI assistant entered their work, issues resolved per hour rose about 15 percent on average. Less experienced agents saw the largest gains. (Brynjolfsson, Li and Raymond, Generative AI at Work, The Quarterly Journal of Economics, 2025)

That is a real productivity gain. It doesn’t tell you what the recovered time should become, whether the remaining work just got harder, or which team receives the extra output next. Those answers belong to leadership.

The technology can complete the task. The organization has to decide what happens because it did.

From Follow the Consequence

What you will be able to do

Inside, you will learn how to

• See what AI is really changing in your organization, beyond the first result
• Decide what the time AI frees up should become
• Show leadership what the business actually gained

Grounded in real operations

Drawn from the work, backed by the research

The book draws on 30 cited research and industry sources. It is grounded in real operating experiences from the author’s career, anonymized. In one, a service desk automated a single task, turning incoming emails into tickets, and within weeks the queue looked flooded.

That was the real change. Demand had not jumped. Visibility had.

From Chapter 9, Measure Past AI

What’s inside: ten chapters follow a change from the first automated task to the whole organization: the work, the people, the authority, the structure and the results. Plus the Introduction, the Conclusion, the Leader’s Toolkit and full endnotes.

Not a book against AI

For leaders who want AI to succeed

This is not an argument against AI. It is not an argument for preserving every job or keeping a person inside every workflow, and it is not another warning that leaders should slow down until every consequence can be predicted.

If AI can perform work better, faster and at lower cost, that matters. Some roles will change. Some will disappear. New ones will emerge. Some AI systems should receive more authority when the evidence supports it.

The goal isn’t to preserve yesterday’s organization. It’s to understand what AI changed well enough to make an intentional decision about what should replace it.