Digital Economy Dispatch #231 -- Does AI Make Us More Productive?

In recent weeks, the focus for many of my discussions on AI adoption has changed. Earlier conversations about the latest AI tools and emerging AI capabilities have been replaced by a deeper, more urgent concern: Are we actually becoming more productive? Addressing this question isn't just an academic issue — it is critical to any use of AI as we look for economic justification for AI implementation across industries.

The Productivity Paradox Revisited

The challenge of demonstrating productivity gains from digital technology isn't new. In 1987, economist Robert Solow famously observed, "You can see the computer age everywhere but in the productivity statistics." This "productivity paradox" has haunted digital transformation efforts for decades. Despite massive investments in information technology, productivity growth in many developed economies has remained stubbornly slow. The Financial Times refers to the UK’s productivity puzzle as “a crisis”.

The extent of this decline is eye-opening. The McKinsey Global Institute reported that productivity growth across advanced economies has declined by nearly 50% since the early 2000s, despite huge increases in computing power, connectivity, and data. Now, as we adopt GenAI and machine learning systems at unprecedented scale and speed, we must ask whether we’ll face similar issues or, in contrast, find ways to define a new phase of productivity growth.

Why Might Productivity Gains Be Elusive?

Is It Too Early?

Historical technology revolutions suggest there's often a significant lag between adoption and productivity gains. Economist Paul David, studying the implementation of electric power in manufacturing, found that productivity benefits took up to 40 years to fully materialize. This delay occurred because realizing benefits required complementary innovations in business processes, organizational structures, and workforce skills.

Similarly, AI may be in its early implementation phase. Organizations are experimenting with use cases, developing capabilities, and learning how to integrate AI into workflows. The most significant productivity gains may only emerge as organizations redesign entire business processes around AI capabilities rather than simply automating existing tasks.

Do We Have the Right Metrics?

Our traditional productivity measures—output per hour worked or per worker—may be inadequate for capturing AI's impact. Knowledge work, where AI shows particular promise, presents measurement challenges that manufacturing productivity metrics don't address.

For instance, work by digital economists like Erik Brynjolfsson raises questions such as how do we measure the productivity of a product manager who uses AI to generate better strategic insights? Or of a data scientist who can now explore more complex models? Traditional metrics might completely miss these improvements in decision quality, innovation capacity, and strategic insight.

Additionally, AI often creates value through improved customer experiences, reduced errors, and enhanced service delivery—benefits that don't always register in conventional productivity statistics but significantly impact organizational performance.

Are We Measuring the Wrong Things?

Perhaps most fundamentally, we may be focusing on the wrong outcomes. The greatest value of AI might not be in traditional productivity but in areas such as:

1. Augmentation rather than automation: AI may be most valuable not when replacing human work but when enhancing human capabilities, enabling workers to perform at higher levels of complexity and creativity.

2. Innovation acceleration: The capacity to experiment faster, iterate more quickly, and discover novel solutions may drive economic value more than incremental efficiency improvements.

3. Resilience and adaptability: AI systems that help organizations respond more effectively to changing conditions create value that isn't captured in short-term productivity metrics.

4. Well-being and engagement: If AI eliminates mundane tasks and allows workers to focus on more meaningful aspects of work, the resulting engagement and retention benefits may exceed direct productivity effects.

Recent research from Stanford and MIT suggests that knowledge workers using AI assistants completed tasks 40% faster while producing higher quality outputs. Yet these impressive results don't always translate to higher-level organizational productivity metrics, suggesting our measurement approach may need recalibration.

AI Productivity: A Multi-dimensional Perspective

A further consideration when examining AI's productivity potential is to differentiate how it impacts distinct organizational groups.

At the individual level, productivity gains are perhaps the easiest to implement and to see immediate returns. Knowledge workers using AI assistants for content creation, analysis, and routine tasks often experience immediate efficiency improvements. These individual productivity boosts represent the "low-hanging fruit" of AI implementation—visible within weeks of adoption. Yet these gains often plateau, suggesting that sustained improvement requires deeper workflow integration.

Team-level productivity follows a more complex trajectory. As AI becomes embedded in collaborative processes, teams typically navigate a period of adaptation where productivity may temporarily decline before improving. However, as teams develop collective expertise in AI collaboration, they begin to realize synergistic benefits unavailable to individual users alone. This transformation typically unfolds over months rather than weeks, requiring deliberate reimagining of collaborative processes.

Organizational productivity represents the most profound but delayed transformation. Here, AI's impact becomes systemic, affecting entire business functions and value chains. Organizations must reconfigure structures, processes, and incentive systems to fully capitalize on AI capabilities. The productivity payoff may take years to fully materialize but ultimately represents the most sustainable competitive advantage.

The interplay between these dimensions creates an important dynamic: early individual productivity gains may create momentum for adoption, but realizing transformative value requires patience and sustained investment in team and organizational capabilities. Digital leaders who understand this multi-dimensional perspective can set appropriate expectations while maintaining stakeholder confidence throughout the complex journey of AI-enabled transformation.

Guiding Principles for Digital Leaders

Understanding these economic concerns with digital technology adoption is critical for AI. As organizations adopt AI at scale, I have seen that digital leaders obtain the clearest return on their AI investments when they consider the following 5 approaches to maximize productive outcomes.

1. Design for Transformation, Not Optimization

The greatest productivity gains come from reimagining work processes rather than automating existing ones. Examine entire workflows and consider how AI fundamentally changes what's possible, rather than simply making current processes more efficient.

2. Invest in Complementary Assets

Technical infrastructure alone won't deliver productivity gains. Invest simultaneously in:

• Skills development and new ways of working.

• Data governance and quality improvement.

• Process redesign and organizational changes.

• Change management and adoption support.

3. Develop AI-Specific Productivity Metrics

Create measurement frameworks that capture AI's unique impacts. These might include:

• Time saved on routine tasks redirected to higher-value work.

• Decision quality improvements.

• Innovation velocity and success rates.

• Error reduction and quality improvements.

• Customer experience enhancements.

4. Take a Portfolio Approach

Balance quick wins that deliver immediate productivity gains with transformative initiatives that may take longer to mature but offer step-change improvements. Track both short and long-term measures of success.

5. Focus on Human-AI Collaboration

The most productive implementations typically don't replace humans but create effective human-AI partnerships. Design systems that leverage the complementary strengths of both human and artificial intelligence.

Reassessing AI Productivity

By expanding our understanding of productivity beyond traditional metrics and designing implementation strategies that address organizational and human factors alongside technical concerns, we can learn from the productivity paradox of previous technology revolutions.

The productivity impact of AI will not follow a linear path, with high costs of initial experimentation followed by more substantial gains as complementary innovations emerge. Digital leaders should ensure they set up these expectations while maintaining focus on productivity as a key outcome.

The question isn't simply whether AI makes us more productive—it's whether we're creating the conditions for AI to fundamentally transform how we create value. That transformation, when it comes, will indeed show up not just in the productivity statistics, but also in our attitudes to adopting AI-at-Scale.