Why the current AI wave is a different kind of change - and what leaders get wrong by treating it as a sequel.
Bernard Sin / Claire Ng · 2 September 2026
Boards are treating the AI wave like a familiar sequel: we did digital transformation; we will do AI transformation. That framing is the first mistake. Digital transformation digitized processes and channels; AI transformation changes how decisions get made, how knowledge work is performed, and how the organization learns. Conflating the two is why so much AI investment is producing activity instead of outcomes.
What Digital Transformation Actually Changed
The term “digital transformation” entered the mainstream vocabulary with the internet boom of the 1990s, as e-commerce and online banking moved long-offline activities onto digital channels (SAS). It matured into a board-level strategic initiative in the mid-2010s, as mobile, cloud, and analytics reshaped how businesses and customers interact, and it accelerated sharply during the 2020 pandemic, when work, commerce, and everyday activity shifted online almost overnight (CIO.com). Its essence, as McKinsey puts it, was the fundamental rewiring of how an organization operates, using technology at scale to improve customer experience and lower cost (McKinsey).
The good: organizations that completed genuine DX have connected data, cloud platforms, and digital channels that would have taken years to build otherwise, and a workforce more comfortable with continuous change. The bad: many DX programs delivered systems without delivering value. McKinsey found that large companies globally captured, on average, only 31% of the expected revenue lift and 25% of the expected cost savings from their digital and AI transformations (McKinsey). Most “transformations” were really modernization projects dressed up in transformation language, measured by what was deployed rather than by what changed.
What AI Transformation Changes Differently
The two share a word and a change-management obligation, but they operate on different substrates. Digital systems are deterministic: the same input produces the same output, and humans direct every decision. AI systems are probabilistic: they reason over data, generate outputs, and increasingly act without a human decision at every step. This is the misinterpretation trap. Leaders hear “transformation” and reach for the old playbook, treating AI as a tool to be deployed rather than a capability to be built. But deploying AI tools is adoption; redesigning how the business decides, learns, and works around AI is transformation.
That shift sounds technical, but it is strategic. In DX, data was a byproduct to be stored and reported. In AI transformation, data is the primary fuel; it determines how systems behave, not just what they can access. In DX, technology executed rules that humans designed. In AI transformation, systems recommend, predict, and sometimes decide, while humans supervise and remain accountable. Digital transformation changed the systems organizations use; AI transformation changes the work organizations do. The risk is no longer slow adoption; it is scaling the wrong operating model.
The timing explains the confusion. AI adoption hovered around 50% of organizations for roughly six years, then surged after the launch of generative AI tools in late 2022. By 2024 it had jumped to 72%, and the latest surveys put regular AI use near 88% of organizations (McKinsey State of AI). Progress is uneven by sector: knowledge-intensive and professional-services functions are moving fastest, while more regulated or asset-heavy environments scale more cautiously (McKinsey). Gartner projects that more than 80% of enterprises will have used generative AI by 2026, up from less than 5% in 2023 — the fastest adoption ramp of any enterprise technology in memory (Gartner). Yet over 80% report no tangible impact on earnings, and only a fraction has scaled AI across the enterprise (McKinsey). Adoption is near-universal; value capture is not. The gap is operating model, not technology.
Three Starting Points, Three Playbooks
This is where the “digital transformation 2.0” framing does the most damage, because starting position changes the playbook entirely.
Mid-transformation, now disrupted. These organizations are years into a DX program, and the ground is moving under them. The instinct is to bolt AI onto the program in flight. The discipline is to ask where AI changes the end-state of the transformation itself. Don’t finish building a process that AI is about to make obsolete.
Never transformed, now jumping to AI. The most tempted, and the most at risk. AI is not a shortcut past the foundations of clean data, connected systems, and digital operating habits. Organizations that skipped DX cannot buy AI maturity off the shelf; they will be building the foundation and the intelligence layer simultaneously, which is harder, not easier.
Completed transformation, facing another wave. These organizations declared DX done and are now realizing the goalposts moved. The trap is complacency, believing the rewiring is finished. The opportunity is that they have the foundation to lead. The work is recognizing that “done” was never the right mental model for something that should compound.
What Carries Over and What Must Change
The lessons from DX that do transfer: start with the business problem, not the technology; technology without operating-model change produces zero durable value; adoption is the real transformation, not deployment; talent and culture are the constraint, not the software.
What must change: success metrics shift from platform adoption (users migrated, processes moved) to decision quality (are people making better decisions, generating novel insight, solving previously unsolvable problems?). Governance shifts from “did the system work?” to “do we trust the output, and who is accountable when it is wrong?” And the timeline shifts from defined phases with an end state to continuous change. An AI-native organization is never done; it gets better at using AI over time. That compounding is the real prize — and it is the one thing a competitor cannot copy by buying the same software.
Key Takeaways
AI transformation is not DX with new tools. It changes decision-making and the nature of work, not just the speed of processes.
Starting position determines strategy. Mid-transformation, never transformed, and “completed” organizations need different playbooks, not the same roadmap.
Adoption is near-universal; value capture is not. Roughly nine in ten organizations now use AI, yet over 80% report no tangible impact on earnings. The gap is operating model, not technology.
Don’t finish building what AI will obsolete. For mid-transformation organizations, re-examine the end-state before adding AI.
You cannot skip the foundation. AI without clean data and connected systems adds cost, risk, and dependency.
Measure decision quality, not deployment. The winners will not be the companies with the most tools, but the clearest transformation logic.
Sources: SAS · CIO.com · McKinsey — Rewired for Value · McKinsey — State of AI · Gartner