SAP Concur says UK data woes stem from culture not tech

SAP Concur has published research suggesting that UK business data problems stem more from organisational culture than from technology. The findings are based on a survey of UK chief executives and finance leaders, with additional input from IT leaders across several markets.

One of the clearest findings was how much manual intervention is still needed before data can be used in business decisions. All 110 UK CEOs and finance leaders surveyed said they must correct flawed datasets before acting on them.

Respondents identified inconsistent definitions, weak data culture and skills, and unclear ownership as the main reasons for those corrections. The figures suggest these problems are embedded in day-to-day processes rather than confined to isolated system failures.

The research also linked those weaknesses to financial work and the use of artificial intelligence in areas such as forecasting and fraud detection. More than half of respondents said strong data foundations improve returns on AI investment, while 57% said poor data quality and integration reduce them.

Manual Fixes

The survey found that 56% of respondents blamed inconsistent data definitions for the need to correct datasets. Another 42% cited a lack of ownership, while 32% pointed to weak data culture or staff skills.

That pattern matters for finance teams because flawed inputs can distort reporting, planning and cost control. When staff must routinely adjust records before analysis, the process becomes slower and more dependent on individual judgment.

Cassie Petrie, managing director for SMB EMEA at SAP Concur, said the findings challenge a common assumption about the source of data errors. "We often hear that organisations are looking for that technological 'silver bullet' to solve their problems, but our research tells us it's often the human foundation that needs work," Petrie said.

AI Friction

The results also show a gap between recognising the value of data and maintaining consistent standards for it. While 50% of respondents said strong data foundations had helped increase returns on AI investments, many also reported that poor quality and integration were holding projects back.

For IT leaders, this was a direct obstacle to the wider use of AI. Some 56% said data quality is a key pain point for AI adoption, affecting projects including fraud detection and cash flow forecasting.

The findings point to a basic problem in how organisations define good data. One team may see data as adequate because it is available, while another may judge it by whether it is accurate, connected across systems and ready to support decisions.

That inconsistency can make it harder to move AI projects from trial to routine use. If standards differ by department, models may be trained on incomplete or uneven information, reducing confidence in the output.

Ownership Disputes

The research also highlighted a divide between finance and IT over who owns data governance and who is blamed when errors emerge. Nearly three-quarters of IT leaders (72%) said data governance is a shared responsibility with finance.

Yet 80% of IT leaders said only the finance team is held accountable when errors occur. No respondents said that only IT is held responsible.

That mismatch suggests many businesses have not set clear rules on ownership, even as they expect data to support more automated processes and AI tools. Without shared accountability, organisations may keep fixing errors without addressing their underlying cause.

Skills were another source of tension. IT leaders identified the main gaps within finance as AI and machine learning knowledge (cited by 80%), followed by data literacy and analytics (73%) and data governance and quality (47%).

These figures suggest that even if finance teams can meet current reporting needs, some technology leaders doubt they are prepared for more data-intensive work. That could become more significant as companies try to embed AI into routine finance operations.

The wider study covered 605 finance leaders, 190 chief executive officers or managing directors, and 180 IT leaders across the United Kingdom and other major markets. "To avoid adding to the 95% of AI initiatives that fail globally, we have to address the culture around data and turn it into a shared strategic asset, not a burden to be passed between different teams," Petrie said.