Old Problems, New Tools: Why Operational Value Isn't Keeping Pace With Innovation
Learn how industrial organisations can turn AI and advanced analytics into real operational value by reducing complexity, improving decision-making precision, and improving safety, reliability, and productivity.
Why isn’t operational value keeping pace with innovation
Investment in AI is moving at extraordinary speed. Gartner expects worldwide spending to reach $2.6 trillion in 2026, up 47% in a year.1 Yet the same forecast points to a more complicated picture: organisations continue to favour tactical applications that deliver incremental gains in efficiency and productivity, while CIOs struggle to demonstrate tangible business outcomes from their AI investments.
Evidence from manufacturing sharpens the tension. Research published by MIT in 2025, drawing on US Census Bureau data covering tens of thousands of manufacturing firms, found that productivity fell by 1.33 percentage points after companies adopted AI, even after controlling for factors including company size, age, capital and IT infrastructure. Over time, AI adopters went on to outperform non-adopters in productivity and market share. Those gains emerged alongside changes in processes, data infrastructure, workforce capabilities and ways of working.2
The gap between adoption and value is where the harder work begins. New capabilities have to change something in the operation: how a problem is understood, where attention is directed, how decisions are made or how work gets done. The graveyard of pilots familiar to many organisations is one consequence of failing to make that translation: technologies can demonstrate technical potential without delivering sustained operational value.
"AI is one part of an expanding set of tools that includes richer operational data, advanced analytics, external benchmarks and connected technologies."
In generative AI, for example, Gartner reported that at least half of projects had been abandoned after proof of concept by the end of 2025, citing poor data quality, inadequate risk controls, escalating costs and unclear business value.3
For industrial companies, that raises a more consequential question: what can these new capabilities contribute to the operational problems that have proved hardest to shift? AI is one part of an expanding set of tools that includes richer operational data, advanced analytics, external benchmarks and connected technologies.
Those problems are not difficult to find: serious injuries and fatalities, recurring reliability losses, throughput variability, energy and material losses, and productivity that becomes harder to improve without adding further processes, systems and complexity. These are familiar problems. What is different is the evidence, analytical capability and breadth of reference now available to understand them, and the possibility of identifying interventions that were previously difficult to see.
1. Gartner. “Gartner Forecasts Worldwide AI Spending to Grow 47% in 2026.” Gartner, 19 May 2026.
2. Burnham, Kristin. “The ‘Productivity Paradox’ of AI Adoption in Manufacturing Firms.” MIT Sloan, 9 July 2025.
3. Chandrasekaran, Arun. “Why 50% of GenAI Projects Fail — And How to Beat the Odds.” Gartner, 26 January 2026. https://www.gartner.com/en/articles/genai-project-failure
Conclusion
The capabilities available to industrial organisations are changing quickly. The leadership challenge is to translate that progress into a higher level of ambition for the operation.
- Start with the operational problems where the value at stake is material and progress has proved hardest to sustain.
Serious risk, recurring reliability losses, persistent variability and productivity constraints should determine where innovation is applied, rather than the availability of a particular technology. - Put accountability for the outcome with the business leader who owns it.
Define the operational result that needs to change, the value at stake and how success will be measured before deciding which capabilities to deploy. Performance, adoption, risk and economics provide a stronger test of value than technical performance alone. - Use new capabilities to make the operation more precise, not more complex.
Connect data across organisational silos, test internal experience against broader benchmarks and use analytics and operational judgement to identify where attention and intervention can have the greatest effect. Be equally prepared to remove activity that is consuming capacity without materially improving the outcome. - Design the change into the operation from the outset.
New insight creates value only when it changes decisions, routines or how work gets done. That requires the people closest to the operation to understand and trust the intervention, with safety, cyber, data and human oversight built in from the beginning.
Competitive advantage will increasingly belong to organisations that can harness innovation to achieve levels of operational performance others cannot.