Despite widespread adoption of artificial intelligence tools, organizations have yet to see the productivity gains forecasters predicted, according to a report published by the University of Manchester's Productivity Institute. The study, titled "The AI Ambition Gap" and commissioned by IT services firm Bytes Technology Group, examines why growing use of AI hasn't delivered expected results. The researchers say closing this gap will require patience and strategic investments beyond the technology itself.
The report points to several concrete examples of the disconnect between task-level improvements and organization-wide outcomes. Software developers using GitHub Copilot finished roughly 26 percent more tasks, while professionals using ChatGPT for routine writing and analysis cut task completion time by 40 percent and boosted output quality by 18 percent. Yet a separate study of AI development tools found that coding activity jumped approximately 240 percent, while actual software releases grew only about 30 percent—a gap the researchers attribute to humans still needing to review and test their work before deployment. The report notes that Britain has the highest proportion of workers in jobs with both high AI exposure and "high complementarity" compared with five other nations studied: the United States, Brazil, Colombia, India and South Africa.
The report argues that high task exposure to AI doesn't automatically make a worker more productive, since jobs typically combine automatable activities with tasks requiring judgment, coordination, client interaction and critical thinking. As AI removes or speeds up some activities, those complementary human tasks become the constraint on time savings. According to lead author Bart van Ark, professor at The Productivity Institute, "AI's productivity potential becomes real when faster tasks lead to better performance across the business." The researchers state that AI may partly displace labor from existing tasks while also creating new tasks where workers retain a comparative advantage.
One explanation for disappointing results is that companies have invested heavily upfront on retraining staff, reorganizing workflows and other adjustments—costs that appear immediately while payback comes later. The report cites recent OECD estimates suggesting high-exposure economies like the UK and US could see annual labor productivity gains between 0.4 and 1.3 percentage points, with even an annual uplift around 0.5 to 0.7 percentage points representing a significant shift relative to Britain's stalled productivity performance over the past decade if sustained. This is welcome news for the UK government, which outlined a strategy last year to use AI for economic recovery. Britain's AI sector is relatively large by European standards, and the country's substantial financial and professional business service sectors—characterized by a high share of cognitive occupations—position it well for AI-driven productivity gains.
The researchers recommend that companies focus on high-value use cases linked to strategy, customer value and operational bottlenecks, plus treat AI as workplace redesign rather than simple technology deployment. The Productivity Institute lists several "mutually reinforcing enablers" including management setting strategic direction, strengthening worker capabilities through greater knowledge and data literacy, implementing AI governance, and focusing on applications that can create new products or solve previously intractable problems. The report suggests policymakers could help businesses by assisting in identifying relevant opportunities to reduce adoption risks and uncertainty. The conclusion: companies must maintain faith and continue investing in AI, and eventually the benefits will materialize if organizations do the right things. The waiting game may test executive patience, but the alternative—abandoning initiatives before they mature—risks leaving gains on the table for competitors willing to stay the course.