The Productivity J-Curve best explains AI’s productivity paradox. The investments across different categories beyond technology initially depress measurable output. Rather than making the necessary investments across these categories, CEOs of legacy corporations often fall into the Delegation Trap, treating AI as a tactical IT upgrade. This misallocation triggers the AI Stakeholder Squeeze, trapping companies between demands for immediate ROI and employee resistance, anxiety, and other important forces. To escape this trap and survive the legacy penalty, CEOs must retain the strategic mandate, avoid delegation, and build a Structural Bridge to safely guide the enterprise from the J-Curve’s investment phase to its harvesting phase.
AI’s Productivity Paradox
It is easy to believe that AI is facing a paradox. While today’s AI capabilities are large and the amount invested in the field unprecedented, meaningful improvements to enterprise productivity to date have not been proportional. This paradox is not unique to AI. It is observed during the adoption and deployment of General-Purpose Technologies (GPTs). The Productivity J-Curve explains it.
The GPT paradox lasts longer than many expect or hope because GPTs diffuse slowly. Their diffusion is constrained by how quickly organizations can adapt their human and institutional structures to them. For example, electrification took decades to deliver productivity gains because factory architecture and workplace organization had to be completely redesigned. Internet adoption was faster, yet it still took a couple of decades for legacy companies to restructure their distribution channels and supply chains.
A GPT, like AI, delivers productivity gains only after organizations make the necessary structural adjustments. Many of these adjustments involve radical transformations of the enterprise’s processes, models, and relations. To maximize the returns during the J-Curve’s harvesting period, it is necessary to make these adjustments during the J-Curve’s investment period.
The Executive Reflex
Our research is starting to show that the AI Stakeholder Squeeze first appears during the J-Curve’s investment phase. During this period, various forces (shown in Figure 1; updated from an earlier version) emerge from within and outside the enterprise. For example, fear of missing out and pressure from the company’s directors and shareholders impact the CEO’s decisions about the enterprise’s AI adoption. Employees responding to the CEO’s pressure to adopt AI tools create unforeseen financial pressures as they frequently exceed the corporation’s annual AI budget within a few months.
While the Squeeze extends into the J-Curve’s harvesting phase, it is most intense during the investment phase due to the number and strength of forces present.
Early data shows that the CEOs of many legacy corporations use macroeconomic headwinds, geopolitical challenges, and various tactical issues as reasons to delegate AI deployments to an executive or an AI committee. This executive reflex shows that they approach AI not as a GPT and a strategic enterprise challenge, but as a tactical software upgrade. The result is the Pilot Purgatory, characterized by localized experimentation without workflow integration and enterprise-wide deployments.
Instead of the exhibited reflex, the CEO must take full responsibility for the adoption and deployment of AI by the corporation. The CEO must ensure that the corporation undertakes the required transformations, completes them successfully, and the workforce embraces the results. The CEOs of legacy corporations, which typically lack digital DNA, feel the AI Stakeholder Squeeze.
The Productivity J-Curve
Succeeding with AI requires ongoing investments in many areas across the enterprise. These investments must continue beyond the J-Curve’s investment phase. In addition to AI technologies, enterprises should be investing in
IT infrastructure.
Employee retraining.
New process creation and existing process re-engineering.
Organizational redesign. This includes adjustments to the workforce composition, creation of new organizational structures, and use of hybrid workforces consisting of both humans and AI agents.
In fact, according to the $1:$9 ratio rule, for every $1 spent on measurable AI technology, successful companies should expect to spend as much as $9 in unmeasured intangible supporting capital. However, today, the legacy corporations, particularly those whose CEOs delegate the AI deployment, invest only in the AI technology and ignore the other areas.
The Legacy Penalty
When adopting a GPT, legacy enterprises suffer a severe productivity drop during the J-Curve’s investment phase because they must actively unlearn uncompetitive management practices and address technical debt. Abandoning established structured management practices accounts for roughly one-third of the initial productivity loss experienced by legacy companies adopting AI. On the other hand, digital native companies have no legacy friction to unlearn when adopting such technologies. Consequently, digital natives may need to invest in fewer categories than legacy companies, allowing them during the harvesting phase to reach a steeper hypergrowth slope and sustain it for a longer period than legacy corporations.
Technology-specific committees and Centers of Excellence permanently trap the legacy enterprise at the bottom of the J-Curve by failing to execute the required organizational change. Legacy corporations require a structural bridge. This is a dedicated, cross-functional organizational vehicle that enables the corporation to concurrently explore new models and exploit established ones. It is equipped with P&L authority and CEO override capabilities, designed to bypass legacy friction and build new AI-native workflows without requiring consensus from the broader matrix organization. Legacy companies that successfully adopted the Internet isolated cross-functional talent into such structural bridges, e.g., Walmart Labs, Disney Digital Network, providing them with their own P&L, technology, and operational authority, and allowing them to bypass the corporate immune system to build the foundation for the harvesting period.
The CEO Squeeze vs the Employee Squeeze
During the J-Curve’s investment phase, legacy corporations tend to misallocate their resources. Organizations get stuck at the bottom of the J-Curve because they treat AI as a plug-and-play software upgrade. Legacy company CEOs often fail to view AI as a strategic issue because they fall into the Cognitive Traditionalist Trap. They assume that because AI utilizes information technology, it belongs under the purview of the CTO. Furthermore, the immense pressure for short-term quantifiable results forces CEOs to prioritize immediate financial metrics over the required long-term, intangible, hard-to-measure investments. These investments place AI in the center of the corporation’s core processes.
Haphazard application of AI to business processes results in employee resistance, confusion, and weaker-than-expected performance. A May 2026 survey conducted by Morgan Stanley provided the data that confirms the anxiety and disruption that employees feel. It revealed that over the last twelve months, in sectors with heavy AI exposure, twelve percent of jobs were eliminated, and an additional fifteen percent of roles were not backfilled. It is particularly present among software engineers. Detroit automakers cut over twenty thousand US-based salaried positions, nineteen percent of their combined workforces, due to the evolving technological changes in the industry, including AI. This empirical data shows that AI adoption in legacy enterprises could fuel active employee resistance because the workforce uses disconnected AI tools without re-imagined workflows.
The CEO Call to Action
The legacy corporation’s CEO must:
Endure the Squeeze and avoid delegating the corporation’s AI adoption and deployment. The CEO must retain ownership of the organizational architecture.
Create a Structural Bridge to facilitate the corporation’s investments and its transition to the J-Curve’s harvesting phase. Such a bridge helps overcome the paradox and lessen the effects of the AI Squeeze. Without such a bridge, legacy corporations risk operational collapse.
Lead the enterprise’s investments and transformations. Determine which core processes AI will need to transform entirely, and which peripheral processes AI only needs to enhance. Determine whether process transformation will require structural changes to the enterprise’s architecture.
Prepare the corporation (employees, shareholders, and partners) for the transformations and continuously reinforce the harvesting phase’s benefits.
