UK enterprises are increasing AI investment, but most are not yet seeing productivity gains at scale, according to Snowflake. UK organizations are heavily funding artificial intelligence, they are currently struggling to achieve significant productivity improvements at scale. Although a vast majority of businesses plan to maintain or grow their AI budgets, progress is being hindered by internal structural issues such as talent shortages and poor data quality rather than technical failures.
The study, conducted by YouGov and based on responses from 500 senior decision-makers across large UK enterprises, highlights a growing gap between AI ambition and execution at scale. While 45% of organizations report early or use-case-specific productivity gains, only 23% say they have achieved gains at scale.
The Investment Paradox
The study reveals a high level of optimism regarding AI’s potential. 99% of UK organizations expect to maintain or increase their AI investment over the next 12 to 24 months, with only 1% planning a decrease. This surge in spending aligns with government ambitions, such as the 2025 AI Opportunities Action Plan, which aims to boost the UK economy by £47 billion annually and increase national productivity by up to 1.5% each year. https://www.gov.uk/government/publications/ai-opportunities-action-plan/ai-opportunities-action-plan
However, the delivery of these gains is currently varied:
- 45% of organizations report early or modest productivity gains.
- 23% have successfully achieved productivity improvements at scale.
- 40% of leaders expect it will take two years or more before AI materially improves their productivity.
Structural Barriers vs. Technological Promise
According to the research, the primary barriers are not technological. Instead, organizations point to internal challenges such as skills shortages, poor data quality, organizational silos and a lack of clear strategic direction.
Technology itself was cited as a barrier by just 19% of respondents, indicating that foundational organizational issues are slowing progress more than platform capabilities.
Governance also remains fragmented. While executive leadership often oversees investment decisions, responsibility for AI strategy and implementation is distributed across multiple functions, limiting accountability and slowing decision-making.
Dr. Fabian Stephany, an economist at the University of Oxford, said that this delay is expected, as “technological breakthroughs rarely translate immediately into productivity improvements,” because organizations require time to adapt their workflows and governance.
Analysis
What This Means for ERP Insiders
AI scaling will hinge on operational integration, not pilots. Enterprises will need to embed AI directly into ERP processes such as finance, supply chain, and procurement to move from isolated gains to enterprise-wide productivity impact.
Cost Reduction Leads AI Success Metrics
When evaluating the impact of AI, UK organizations are prioritizing operational efficiency over growth outcomes.
Nearly half of respondents, 44%, identified cost reduction as the most important measure of AI success, compared with 26% who cited revenue growth. This indicates that many organizations are approaching AI primarily as a tool for optimization rather than transformation.
At the same time, timelines for measurable impact remain extended. Around 40% of organizations expect AI to take two years or more to deliver meaningful productivity improvements.
Analysis
What This Means for ERP Insiders
Efficiency-led AI strategies may cap long-term business impact. While cost reduction remains a primary driver, organizations that expand AI use cases into revenue growth, customer experience, and decision intelligence will be better positioned to realize sustained competitive advantage.
Confidence Tempered by Governance and Risk Concerns
The research highlights a cautious approach among senior leaders when it comes to scaling AI initiatives.
Only 24% of organizations say their AI initiatives are prioritized using a structured framework aligned with business objectives. In parallel, concerns around ethics, safety and reliability continue to shape adoption decisions.
Around 60% of respondents say these concerns influence how they deploy and scale AI, particularly in regulated industries.
This cautious stance is especially visible in the public sector, where organizations demonstrate strong governance but expect slower timelines for productivity gains due to risk management requirements.
Analysis
What This Means for ERP Insiders
Data and governance maturity will define execution speed. Organizations with fragmented data landscapes and unclear ownership structures will struggle to scale AI, making data harmonization and governance frameworks critical enablers of value realization.
Industry Differences Reflect Uneven Maturity
The research points to varying levels of AI maturity across industries:
- Financial services organizations show stronger governance and strategic alignment but face regulatory and reputational constraints that slow scaling.
- Manufacturing firms express confidence in long-term productivity gains but anticipate slower returns due to skills gaps and integration complexity.
- Retail organizations lag in both confidence and execution, with AI initiatives often limited to isolated use cases due to persistent data quality challenges.
- Public sector organizations demonstrate the highest levels of risk awareness, with 53% citing safety and reliability of AI outputs as a primary concern and 52% expecting productivity gains to take more than two years. and strategic alignment but face regulatory and reputational constraints that slow scaling.
Closing the Gap Between Ambition and Impact
The findings point to a broader shift in enterprise AI strategies, from experimentation to structured execution.
According to Snowflake, organizations that succeed in scaling AI will need to focus on stronger data foundations, clearer governance models and tighter alignment between AI initiatives and measurable business outcomes.
While investment levels remain high, the report suggests that unlocking AI’s full productivity potential will depend less on new technology and more on organizational readiness.
