'Belief alone is not enough': New study reveals businesses are widely different when it comes to AI adoption — so what's the best course?
45% of UK firms report AI productivity gains, but execution gaps risk leaving many behind.
A new study from data cloud company Snowflake reveals a stark divide in corporate AI adoption, with nearly all businesses (99%) planning to maintain or increase their AI investments over the next 1-2 years. However, execution is failing to match ambition: only one in four organizations (24%) currently use a clear framework to align AI initiatives with measurable business objectives. While 45% of UK firms report small-to-modest productivity gains, and 23% have achieved them at scale, the research identifies internal barriers—not technology—as the primary obstacle. Poor data quality, organizational silos, skills shortages, and a lack of clear leadership and strategy are blocking stronger adoption and ROI.
Snowflake Principal Data Strategist Jennifer Belissent emphasized that 'belief alone is not enough,' noting that real productivity gains require clear ownership, strong data foundations, and strategic alignment. The challenges are not uniform across sectors; financial services companies are constrained by tight regulation, while retail lags primarily due to data issues. Furthermore, ethics and safety concerns are shaping deployment for two-thirds of organizations. The report concludes that unlocking AI's potential depends on getting fundamentals right: governance, data quality, and clear accountability, warning that companies failing to address these execution gaps risk being left behind as the AI landscape evolves.
- 99% of businesses plan to maintain or increase AI spend, but only 24% have a clear strategic framework for it.
- Internal barriers like poor data quality and organizational silos—not technology—are the main blockers to AI ROI.
- Sector-specific challenges exist, with financial services hindered by regulation and retail slowed by data issues.
Why It Matters
For professionals, this highlights that successful AI adoption requires fixing data and governance first, not just buying more tools.