AI Trust Paradox Exposes Gaps in Literacy & Governance
Informatica by Salesforce has published a study highlighting a trust paradox surrounding corporate AI use. The research reveals that, while employee confidence in AI data is rising, there are significant concerns regarding data and AI literacy and inadequate governance.
Study Overview
The report, titled “CDO Insights 2026: Data Governance and the Trust Paradox of Data and AI Literacy Take Center Stage”, is based on a survey of 600 data leaders from Europe and the UK. It includes insights on the adoption of generative AI and agentic AI, along with investment priorities for data management and workforce training.
Across Europe, 68% of respondents indicated their businesses would initiate agentic AI pilots by the end of the first quarter of 2026. In the UK, 61% plan to transition towards becoming an agentic enterprise.
Furthermore, the study reported that 79% of European businesses are expected to adopt generative AI within the same timeframe.
Understanding the Trust Paradox
The findings highlight a significant gap between perceived trust and readiness. While 61% of European data leaders expressed confidence that employees trust most or all of the data used for AI, only 52% of UK data leaders shared this belief.
Simultaneously, 96% of European respondents emphasized the need for enhanced training in AI and data literacy to responsibly utilize AI and its outputs. This includes 82% advocating for more data literacy training and 71% for AI literacy training.
Informatica associated this literacy gap with governance issues, as 77% of European respondents noted their organization’s AI visibility and governance have not kept pace with employee AI technology usage.
Drivers of AI Adoption
The research explored motivations for broader AI adoption within organizations. European respondents identified improving business decision-making and enhancing employee collaboration as the primary drivers, each at 32%. Optimizing internal processes followed closely at 28%, with enhancing customer experience and loyalty at 27%.
When sourcing agentic AI tools, 55% of European respondents plan to purchase vendor-supplied agents, while 44% in the UK anticipate a similar route. Additionally, 45% of European organizations and 55% in the UK intend to develop and manage agents internally, with 21% planning to use no-code or low-code platforms.
Investment Priorities
The study indicates a shift in spending as AI moves from pilot phases to broader deployment. 85% of European organizations expect to increase spending on data and AI management in 2026, with 23% anticipating significant increases.
Data leaders highlighted three key areas for additional funding, each at 44%: upskilling employees to enhance data and AI fluency, improving data privacy and security, and strengthening data and AI governance. In the UK, 49% of organizations plan to invest in enhancing data literacy.
Barriers to Production
Overall, data quality and reliability emerged as persistent barriers. The study found 57% of European data leaders view poor data reliability as a significant challenge in advancing generative AI initiatives from pilot to production. In the UK, this concern rises to 60%.
Concerns also exist regarding AI pilots progressing without addressing underlying reliability issues. 50% of European data leaders expressed serious concern about new AI pilots moving forward without rectifying previous data reliability problems, while 46% in the UK shared similar apprehensions.
For AI agents entering production, 51% of European organizations identified data quality and retrieval as the top challenge, followed by security concerns at 46% and a lack of expertise in agentic AI at 45%.
Respondents also mentioned operational controls and tooling challenges, with 40% citing observability issues in Europe and 43% pointing to insufficient tools for managing AI agents. Additionally, 41% reported a lack of safety guardrails.
In response, organizations are adopting process changes and investing in better data practices. 58% are improving workflows around data and AI, 56% are investing in data and metadata collection and management, 55% are increasing the frequency of data checks, 54% are enhancing investments in data quality, and 52% are hiring or upskilling staff in this area.
Conclusion
The report encapsulates the notion that agentic AI without robust data governance is not true innovation but rather an exposure to risk. Blind trust in AI, devoid of corresponding data and AI literacy, creates a false sense of confidence among organizations. However, early signs of maturity in AI strategy are emerging, as data leaders increasingly recognize the need for investments in data governance and compliance foundations to mitigate risks while maximizing opportunities.
