J Curve and Solow Productivity Paradox are at Work with AI

Investors are going to keep challenging firms to show evidence their heavy artificial intelligence investments really are boosting productivity.

That is going to continue being a tough challenge, as history suggests the real output gains will take some time to develop.

So AI "productivity," or the "lack of quantifiable gains," are currently the most significant contemporary case of the Solow productivity paradox. 

In 1987, Nobel laureate Robert Solow famously remarked, "You can see the computer age everywhere but in the productivity statistics."

Recent research suggests productivity might actually decline for a time as firms deploy AI. 

The reason is the J curve. 

“We find causal evidence of J-curve-shaped returns, where short-term performance losses precede longer-term gains,” say economists Kristina McElheran; Mu-Jeung Yang; Zachary Kroff and Erik Brynjolfsson. “Consistent with costly adjustment taking place within core production processes, industrial AI use increases work-in-progress inventory, investment in industrial robots, and labor shedding,

while harming productivity and profitability in the short run.”

In other words, it takes time for enterprises to retool their business processes for the new technologies. And the more profound the innovations, perhaps the longer it takes to integrate those tools. 

Also, much of the reported AI adoption is horizontal rather than vertical; personal rather than systematic. In other words, individuals might be using chatbots, but workflows have yet to be transformed. 

So “personal productivity” has not yet been matched by an applied transformation of key work processes. And personal productivity gains are hard to measure, in terms of impact on firm performance. 

Agentic AI should help, as they can affect complex business processes. 

source: Forbe

Many have noted that  U.S. labor productivity significantly slowed in the 1970s and 1980s, despite rapid information technology investment. 

Then starting in the mid 1990s a decade of faster growth returned arguably because business process re-engineering had taken place.

A similar productivity paradox surrounds AI. As explained by economists Erik Brynjolfsson, Daniel Rock, and Chad Syverson in a 2017 working paper, AI and the Modern Productivity Paradox,” the paradox is primarily due to the time lag between technology advances and their impact on the economy. 

While technologies may advance rapidly, humans and our institutions change slowly. 

Moreover, the more transformative the technologies, the longer it takes for them to be embraced by companies and industries across the economy.

Translating technological advances into productivity gains requires major transformations, and therefore time.

Today, we see a "Modern AI Paradox": while Large Language Models (LLMs) and Generative AI are ubiquitous in headlines and corporate pilots, global aggregate productivity growth  remains sluggish.

Economists like Erik Brynjolfsson argue that the paradox isn't a failure of the technology, but a timing and structural issue. He identifies four main reasons for this lag:

1. Mismeasurement: AI often improves quality, variety, or speed in ways that traditional GDP (which tracks "units produced") fails to capture.

2. Redistribution: AI may be used for "rent-seeking" (competing for market share) rather than increasing total industry output.

3. Implementation Lags: Significant "General Purpose Technologies" (like electricity or the steam engine) require decades of organizational restructuring before they move the needle.

4. Mismanagement: Companies often use AI to automate old processes rather than inventing new, more efficient business models.

Study	Target Group	Productivity Impact Found	Notes on Enterprise Deployment Gaps
MIT/Stanford (NBER)	Customer Support Agents	14% increase in issues resolved per hour.	High-skilled workers saw less gain; impact was greatest on novices. Enterprises often fail to use AI as a "leveler" for training.
Harvard/BCG (SSRN)	Management Consultants	40% higher quality; 25% faster task completion.	"Jagged Frontier": AI failed spectacularly on certain logic tasks where humans over-relied on it, leading to "falling off the cliff" errors.
Microsoft/GitHub	Software Developers	55% faster at completing coding tasks.	Gains are often eaten by "code bloat" and increased technical debt if not managed by senior architects.
Goldman Sachs Research	Aggregate US Economy	Projected 1.5% annual increase over 10 years.	Real-world adoption is currently hindered by power grid constraints and data center infrastructure delays.
NBER / Brynjolfsson et al.	Generative AI & the "J-Curve"	Initial 0% or negative impact.	The "Productivity J-Curve": Measured productivity dips initially as firms invest in "intangible capital" (retraining, restructuring) before the payoff.

While individual tasks show gains, enterprise-wide productivity often remains flat for several reasons:

• The "Pilot Trap": According to recent Adobe/Business research, 86 percent of IT leaders see potential, but only a fraction have moved beyond "isolated experiments" to organization-wide workflows

• Inertial Workflows: Companies often use AI to "do the old thing faster" (e.g., writing more emails) rather than "doing the right thing" (e.g., eliminating the need for those emails entirely). This results in "Digital Overload"

• The Human Bottleneck: AI can generate a report in seconds, but a human still takes hours to verify, edit, and approve it. Without changing the governance and approval structures, the AI speed gain is neutralized

• Data Fragmentation: Most AI models are effective only if they can access clean, centralized data. Most enterprises still have "siloed" data, leading to AI hallucinations or irrelevant outputs

• Skills Gap: Enterprises frequently treat AI as a "plug-and-play" tool like a calculator, failing to realize it requires a new type of "AI Literacy" to prompt and integrate effectively into complex projects.

None of that will be too comforting for suppliers who must justify their heavy AI capital investment. 

But history suggests the payoff is coming. It just will take some time. It always does.

"Not Seeing AI Productivity" Storyline is Inevitable

It’s inevitable that we will keep seeing headlines, and seeing, hearing and reading stories about how many businesses are not seeing financial returns from their investments in artificial intelligence. 

Important new technologies rarely show up in the bottom line immediately, and the issues are structural.

First of all, business processes have to be recreated to harness the innovations. 

When electricity first entered factories, managers simply replaced their massive central steam engine with one massive electric motor. Productivity didn't move. Only after firms discovered they could put a small motor on every individual machine (the "unit drive") were they able to redesign the factory floor. 

In 2026 companies are using AI to "chat with docs" or "summarize emails" (overlaying tech on old habits) rather than redesigning their entire supply chains. That will take some time. 

Also, firms must retrain workers and staffs. That imposes real costs (time and money) while possibly lowering productivity in the short run as time and effort is diverted to such training. So a "J-curve" of productivity will happen: lower productivity in the near term, with the benefits in the future. .

Then there are measurement issues, such as how to quantify the impact of quality, variety and speed. If an AI helps a legal team finish a contract in two hours instead of 10, but the firm still charges a flat fee, the "productivity" is invisible to the GDP, even though the human cost has plummeted.

Study	Technology Period	Key Finding	Duration of Lag
Paul David (1990)	Electricity (1890–1920)	Factories had to be physically demolished and rebuilt to utilize "unit drive" motors before TFP spiked.	~30–40 Years
Robert Solow (1987)	Computing (1970–1990)	The "Solow Paradox": You could see the computer age everywhere except in the productivity statistics.	~20 Years
Brynjolfsson et al. (2021)	AI & Software (2010–2021)	Formulated the "Productivity J-Curve"; firms must invest in unmeasured "intangible assets" that initially depress earnings.	Ongoing
NBER / Juhász et al. (2020)	19th Century France	Productivity in mechanized spinning was initially lower and more dispersed than hand-spinning due to the need for factory reorganization.	~15–20 Years
Man Group / Bara (2026)	Generative AI (2023–2026)	80% of firms report no macro productivity impact yet, despite task-level gains of 15-55%, due to "workflow friction."	Projected 5-10 Years

In the meantime, leaders will have to try and come up with some quantifiable metrics (directly related or not) to justify the investments. It won’t take too much imagination to realize that headcount reductions are one such way to “show” outcomes, even if AI and headcount are indirectly, loosely or even unrelated in the short term. 

In 1900 the "electricity bubble" looked real to everyone still using steam. By 1920, the steam users were bankrupt. 

So “productivity proxies” must be developed.

The most immediate impact of AI is the compression of time. 

Firms can measure the "distance" between an idea and its execution.Time-to-prototype can show how many days it takes to move from a natural language prompt or requirement to a functional, testable version.

Draft-to-final ratio might be used by marketing and legal firms to measure the time spent on the "first 80 percent" of a task versus the "final 20 percent" of human polishing.

For engineering teams, the metric isn't just "lines of code," but the number of successful pushes to production per developer per week. 

Larger firms might try to assess the reduction in total "human hours" spent in meetings.

Query-to-find latency is a measurement of how long it takes an employee to retrieve a specific piece of internal tribal knowledge. AI should reduce that latency. 

Admin-to-maker ratio tracks whether the percentage of an employee's day spent on "coordination" is shrinking in favor of "creation." 

“Agents” also will need new metrics that quantify AI outcomes as though it were a digital employee rather than a software tool.

Autonomous completion rate is the percentage of workflows that an AI agent initiates and completes without a human "click" or intervention.

Human-in-the-loop friction measures how often an agent has to "hand back" a task to a human because it hit a reasoning wall. A falling HITL rate is a leading indicator of future productivity.

Token efficiency per outcome calculates the cost of AI "thinking" (API/Compute costs) relative to a successful business outcome. 

Business Function	Traditional Metric (Lagging)	AI Proxy Metric (Leading)	Why it Matters
Software Engineering	Lines of Code, Story Points	PR Cycle Time	Measures how fast code is reviewed and merged, not just typed.
Legal, Compliance	Billable Hours	Review Velocity per Page	Shows the acceleration of document ingestion and risk flagging.
Customer Support	First Response Time	Resolution via Zero-Touch	Measures the percentage of issues solved entirely by agents.
R&D	Patents Filed, Products Launched	Iteration Cycles per Quarter	Shows how many "failed fast" experiments the firm can run.
Human Resources	Headcount Growth	Talent Density (Revenue/FTE)	Measures if the firm is scaling output without scaling people.

The “productivity lag” is entirely predictable. So are the storylines about it. Sure, it is a significant practical problem for those firms making the investments. But the “lag” storyline is entirely predictable.