Study: Firms Cut Jobs for Expected AI Gains, Not Actual Value
6 min read
A global study of more than 1,006 C-suite executives across 11 countries and 32 industries has found that while 90% of organisations report receiving value from AI, only a minority have built the measurement and leadership capabilities needed to consistently capture meaningful economic impact. The report — Economic Maturity for Artificial Intelligence: How Organizations Measure and Maximize Value from Artificial Intelligence — was published by the Return on AI Institute, Scaled Agile's applied AI research partner, and co-authored by Thomas H. Davenport, Distinguished Professor at Babson College and fellow of the MIT Initiative on the Digital Economy, alongside Laks Srinivasan, a veteran AI transformation leader.
Workforce Decisions Are Moving Faster Than Realised AI Impact
The report's most striking — and sobering — finding concerns the pace of workforce decisions relative to actual AI-driven productivity evidence. Only 2% of organisations have made large headcount cuts tied to real, demonstrated AI implementation outcomes. Yet nearly 90% have already reduced or frozen hiring in anticipation of future AI productivity gains that have not yet been formally measured or validated.
The pattern is one of expectation-driven workforce restructuring outrunning evidence-based decision-making — organisations acting on the belief that AI will eventually displace significant labour capacity, rather than waiting until that displacement has actually been demonstrated in their operating context. For talent management and talent acquisition leaders, this finding carries significant implications: workforce reductions premised on AI productivity gains that have not yet been captured create organisational capability gaps that can compound over time.
"We've studied technology adoption in organizations for decades. The pattern here is consistent: the technical capabilities arrive before the management systems to harness them. What's different with AI is how many consequential decisions — especially on workforce — organizations are making before those systems catch up."
— Thomas H. Davenport, Distinguished Professor, Babson College; Co-Founder, Return on AI Institute
Measurement Maturity: The 70-Point Gap That Separates AI Leaders from Everyone Else
The single biggest differentiator between organisations achieving strong AI value and those that are not is not the AI technology they deploy — it is how rigorously they measure, aggregate, and report the economic impact of their AI investments. Organisations that formally report AI value to boards or investors achieve high value at an 85% rate. Those that do not measure or report at all achieve it at just 15% — a 70-percentage-point gap driven entirely by measurement discipline rather than technology advantage.
Training Both Employees and Leaders Matters — But Most Organisations Are Missing the Basics
The report identifies a 23-percentage-point advantage in achieving high AI value when both employees and leadership undergo AI training — compared to organisations that invest in one or neither. Yet the baseline is alarmingly low. 58% of organisations have still not trained employees in basic AI productivity and tool use. And 29% acknowledge that their leaders lack the understanding needed to drive AI value creation.
The implication is clear: the returns from AI investment are not primarily a function of which AI products an organisation purchases. They are a function of whether the organisation has built the human infrastructure — employee skills and leadership fluency — to actually put those tools to work in ways that generate measurable economic impact.
"The technology works — 90% of organizations say so. What separates the leaders from everyone else isn't the AI itself. It's whether anyone has the discipline to measure what it's worth and the leadership fluency to act on what they find."
— Laks Srinivasan, Co-Founder and CEO, Return on AI Institute
Generative AI vs Analytical AI: Where Value Is Actually Being Captured
The report provides important context on where AI value is currently being generated versus where the market conversation is loudest. Only 9% of organisations currently identify generative AI as their most valuable AI type — compared with 50% for analytical AI and 40% for rule-based automation, both of which carry decades of operational maturity. The excitement around generative AI in executive communications has significantly outpaced its current share of measured enterprise value.
44% of executives say generative AI is the hardest AI type for which to measure ROI — reflecting that organisations are still learning how to integrate it into workflows and operating models in ways that produce quantifiable outcomes. The emerging category of agentic AI systems is beginning to show strong early returns, with adopters 22% more likely to report achieving a great deal of AI value — suggesting the next wave of enterprise impact may come from AI systems that act, coordinate, and automate work rather than simply generate content.
The US Is Underperforming Its AI Reputation
One of the report's most pointed findings concerns the United States' performance relative to its global perception as the AI leader. Despite being widely regarded as the dominant force in AI technology and investment, only 38% of US organisations report getting a "great deal" of value from AI. That figure is well below Germany, the UK, Australia, Japan, and the UAE — all of which exceed 50%, with virtually identical levels of employee training and AI experience to their US counterparts.
The gap cannot be explained by technology access, investment scale, or workforce AI experience — it is a measurement and management maturity gap. US organisations are generating AI activity at scale; they are not systematically capturing or reporting the economic value that activity produces. The 70-point measurement gap documented elsewhere in the report is likely the primary driver of the US underperformance finding.
Key Takeaways
- 90% of organisations report AI value, but nearly 90% have already frozen or cut hiring in anticipation of future AI gains — while only 2% have made cuts tied to actual, demonstrated AI outcomes.
- Formal AI value reporting to boards drives an 85% high-value achievement rate vs 15% for organisations that don't measure — a 70-percentage-point gap driven entirely by measurement discipline.
- 58% of organisations haven't trained employees in basic AI tools; 29% acknowledge leadership lacks AI fluency — yet combined training delivers a 23-point advantage in achieving high AI value.
- Only 9% identify generative AI as their most valuable AI type (vs 50% for analytical AI); agentic AI adopters are 22% more likely to report high value — signalling where the next wave of enterprise impact is forming.
- Only 38% of US organisations report high AI value — well below Germany, UK, Australia, Japan, and UAE — despite similar training levels, pointing to a measurement and management maturity gap rather than a technology gap.