There’s a striking disconnect between the promise of AI and its measurable impact on UK business productivity, new research reveals.

The study from Qlik, based on insights from 250 business and IT leaders in the UK, shows that despite aggressive investment in AI, many organisations are still struggling to translate pilot projects into productivity gains, and even fewer can tie those gains to clear financial KPIs.

Everywhere, But Inconsistent

Despite widespread deployment, productivity gains for AI continue to be inconsistent. More than half of business and IT leaders report that fewer than 50% of their AI projects have delivered measurable improvements.

Even more concerning, only 11% of respondents say the majority of their initiatives (over 75%) have led to tangible gains.

This is compounded by a lack of financial clarity, with just 51% of organisations using KPIs tied directly to business performance to evaluate AI initiatives. 44% admit their teams’ perceptions of AI’s productivity gains don’t align with the actual outcomes.

“AI adoption is high, but impact remains patchy,” said James Fisher, chief strategy officer at Qlik. “This gap between hype and reality is a wake-up call. Businesses need to focus on measurement, alignment, and building the data infrastructure that enables AI to deliver at scale.”

Leading the Way or Lagging Behind?

Where AI is making a mark, it is largely in technical domains. IT and cybersecurity departments are the standout beneficiaries, with 81% of leaders reporting improvements in these areas. But other business functions aren’t seeing the same uplift.

However, HR and finance remain on the margins of AI success, with 37% and 30% of respondents respectively stating these departments have seen the least tangible benefit. This indicates that the full potential of AI is not being realised across the business, with innovation and investment still concentrated in siloes.

Operational Gap Are the Biggest Barriers

Only a quarter of respondents cited budget as a top-three barrier to AI success. Instead, the biggest challenges are operational, with nearly half (49%) of businesses saying they lack the internal skills to effectively integrate AI with their existing analytics and business intelligence systems.

Other top concerns include incompatible tools and platforms (36%) and a lack of real-time data integration (37%). These infrastructure and capability gaps are limiting the returns on AI investment, even in well-funded organisations.

Recommended reading

• UK Orgs Have High Hopes for Agentic AI ROI
• GenAI Drives 1.7x ROI, New Report Claims
• DataFest | LEGO’s AI Chief on Getting ROI Right

Projects Stuck in Pilot Mode

The research also suggests that businesses are still operating in a proof-of-concept mindset. Nearly a quarter (23%) say that the majority of their AI use cases (75% or more) are still in the experimental phase. A further 11% report that virtually all of their initiatives are still in early-stage pilots.

This signals a major disconnect between perception and reality. Many businesses believe they are ready to scale, but remain stuck at the starting line when it comes to operationalising AI at scale.

Clearer KPIs and Better Tools Needed

Many businesses still struggle to measure AI success effectively. While 89% agree that a unified data strategy is key to assessing ROI, only half say they have the tools to connect AI outputs with business performance.

There’s also strong demand for better tools and transparency. 57% say improved data integration and analytics would help communicate AI’s business value to stakeholders, while 55% want more visibility into how AI models make decisions. Stronger collaboration between IT and other business units (49%) and clearer, outcome-focused KPIs (46%) are also seen as critical to moving from experimentation to impact.

“To realise AI’s full potential, businesses must move beyond experimentation and focus on execution,” said Fisher. “That means scalable tools, integrated strategies and collaboration across every function.”