57% of businesses are piloting or running AI agents, but 8% say their data is production-ready: Survey
Findings from 540+ data leaders and practitioners across 65+ countries show trusted data, business context and governance are lagging behind enterprise AI adoption
***
As enterprises race to move artificial intelligence from experimentation into daily operations, the data foundation holding these systems up is emerging as the critical weak link, according to the latest Modern Data Report: 5 Emerging Trends in Enterprise Data & AI, released by The Modern Data Company.
Based on more than 540 qualified responses from data leaders and practitioners across over 65 countries, the report lays bare a striking mismatch: 57.3% of organisations are already piloting AI agents or running them in production, yet only 8.4% say the data feeding their AI systems is trustworthy enough for production use. In other words, AI adoption is surging ahead while the confidence in the data behind it is barely moving.
Data Quality Is Now The Top Barrier
Data quality and trust have climbed to the number-one barrier blocking AI agents from reaching production, with 76% of respondents ranking it among their top three obstacles. Security and governance concerns follow closely, with 62% placing them in their top three. Only 18% of organisations report having a clear, documented AI accountability framework in place.
The report also notes a persistent confidence gap: 59% of respondents express low or mixed confidence in the trustworthiness of the data behind their AI systems, and even among organisations already running agents in production, only 21.7% describe their data as highly trustworthy.
"The report shows that the challenge is moving from AI experimentation to operating AI reliably at scale. Based on our research, data quality and trust now rank among the biggest barriers to putting AI agents into production, while governance has to extend to how AI uses enterprise data and acts on it. As AI takes on a more active role in business workflows, organisations need to build the context, controls and accountability that allow these systems to operate with confidence," said Saurabh Gupta, President & CEO, The Modern Data Company.
The findings underscore a growing recognition that AI cannot function on raw data alone. 61% of respondents say a reliable context layer — definitions, relationships, policies, lineage and other forms of business meaning — is critical or very important for AI agents. Yet just 16% say their organisation deliberately designs and engineers that context layer as a product. For most, business context remains scattered across tools, documentation and people's knowledge rather than being intentionally built, owned and managed.
Consolidation Shifts From Preference To Action
The report also documents a decisive turn toward platform consolidation. While 77% of respondents at the end of 2025 already strongly believed in the value of a more converged data platform, 47% are now actively consolidating toward fewer platforms and another 17% are evaluating it. Fragmentation still carries a heavy cost: 46% say more than a quarter of their team's time goes toward maintaining or integrating tools rather than delivering data, analytics or AI value. With consolidation underway, the conversation is shifting from whether to consolidate to how to do it without sacrificing flexibility or creating new vendor lock-in.
Governance Expands With AI's Reach
The report finds that AI has not replaced traditional data governance requirements — it has extended them. Organisations still struggle with fundamentals such as lineage (53% had no easy way to trace data lineage at the end of 2025), ownership and access control. Now they must also govern what information AI can use, what actions it can take, which policies apply and who is accountable for the outcomes.
The Leaders Show The Way — With Caveats
Perhaps the most telling signal in the report is the pattern among organisations furthest along with AI. Those already running agents in production are 3.6 times as likely to have deliberately engineered their context layer, and three times as likely to report confidence in their data, compared with organisations interested in agents but yet to begin. They are also substantially more likely to have established clear AI accountability. The report cautions that these relationships do not establish causation, but the pattern is consistent.
Taken together, the report points to a widening gap between AI ambition and data readiness. The core issues at stake are:
- Trust gap: AI is advancing faster than trust in the data behind it, and only 8.4% believe their data is production-ready for AI.
- Context deficit: Most organisations know business context is essential for AI agents, but very few have engineered it deliberately.
- Governance vacuum: Security and governance rank among the top barriers, yet fewer than one in five organisations have a documented AI accountability framework.
- Fragmentation cost: Nearly half of teams lose more than a quarter of their time to tool maintenance and integration, even as consolidation efforts gather pace.
- Operational risk: As AI agents move from assisting people with analysis to acting across enterprise systems and workflows, weak data foundations translate directly into unreliable decisions and actions.
Implications for India
The findings carry particular relevance for Indian enterprises as AI moves from experimentation toward operational use across business functions. "For Indian businesses, the next phase of AI adoption will be about turning AI capability into measurable business outcomes. That requires organisations to build trusted data and business context into the AI foundation from the start," said Sanjoy Roy, Vice President, APAC, Middle East & EU, The Modern Data Company.
The report concludes that the next phase of enterprise AI will be defined less by who has access to the most AI capabilities and more by who can build the enterprise foundation — trusted data, deliberate context, extended governance and clear accountability — required to use them reliably at scale.
The findings are drawn from the third Modern Data Survey, an ongoing research initiative conducted through the Modern Data 101 community, which brings together more than 15,000 members globally across 130+ countries. The interim report primarily reflects responses collected during July and August 2026.
Comments