NORBERT BIEDRZYCKI

The AI Enabled Organization: Why Technology Adoption Is Not Transformation

The AI Enabled Organization: Why Technology Adoption Is Not Transformation

Almost every large organization now uses AI somewhere. McKinsey’s latest State of AI survey, published in November, puts the share of organizations using AI regularly in at least one function at 88 percent. Yet only a small minority, around six percent, qualify as high performers that attribute a meaningful share of their earnings to it. Most companies are still in the pilot and experimentation phase.

An earlier edition of the same survey, from March 2025, contained the finding I consider most important. Of 25 organizational attributes tested, the redesign of workflows had the strongest relationship with an organization’s ability to see EBIT impact from generative AI. Only 21 percent of companies using it had fundamentally redesigned even some of their workflows.

That is the whole story of enterprise AI in two numbers. Almost everyone adopted. Very few transformed.

Adoption is the easy part

Buying licenses, launching a copilot, running a hackathon and building a dozen pilots is now standard practice. It produces visible activity, internal enthusiasm and good slides for the board.

It rarely produces significant value. When AI is added to an existing process, it makes individual steps faster. But the process, with its handovers, approvals and queues, stays the same. The bottleneck moves rather than disappears. A report is drafted in ten minutes and then waits five days for a sign off.

The largest gains come when the entire workflow is reconsidered. Which steps exist only because work used to be manual? Which approvals were a substitute for information that is now available instantly? Which roles exist to move data between systems?

The organizational questions

That is why AI is ultimately an organizational topic rather than a technological one. Redesigning work around it forces questions that most companies have avoided.

Roles. What does a financial analyst, a claims handler or a key account manager do when a large share of their current tasks can be executed by a system?

Structures. How many management layers are needed when information no longer has to be aggregated and reported manually?

Decision rights. Who may let an AI system act, within what limits, and who answers when it is wrong?

Expertise. How is human expertise developed and preserved when the routine work that used to build it is automated?

Measuring the right things

Many leadership teams still report AI progress in terms of adoption: number of users, number of use cases, number of agents deployed. Those are useful as early indicators. They are not results.

The question boards should ask is the one I raised about transformation ROI in 2024. What has changed in productivity, margin, revenue or cycle time, and can we trace it to specific redesigned processes? If nobody can answer, the company has adopted AI. It has not yet become AI enabled.

In my work with PE portfolio companies, the distinction is visible quickly. The companies that create value start with two or three processes that matter to the P&L, redesign them end to end, and only then scale. The companies that struggle start with tools.

Leadership question: Are we using AI to make existing organizations more efficient, or are we building fundamentally different organizations?

Read next

  1. Beyond Efficiency: The Next Transformation Frontier Is Business Model Reinvention

    Transformation and Leadership series, written in 2026

  2. From Managing People to Orchestrating Human and AI Teams: The New Leadership Challenge

    Transformation and Leadership series, written in 2026

  3. Transformation and Leadership 2020–2026: about this series