Gartner released one of its more useful research findings last week. And the headline number deserves careful reading (rather than a quick skim).
Organisations that report successful AI initiatives invest up to four times more, as a percentage of revenue, in foundational areas (data quality, governance, AI-ready people, and change management) compared to those that experience poor outcomes from AI.
Sit with that ratio for a moment. Four times more. Not in models. Not in accelerators. In foundations.
The survey of 353 data and analytics and AI leaders from November through December 2025 produced a second number that should be equally uncomfortable: only 39% of technology leaders are confident that their enterprise’s current AI investments will have a positive impact on financial performance.
More than six in ten are not confident.
This is not a fringe finding, it is the majority view among the people closest to the investment.
The six shifts Gartner identifies for what it calls “AI-first D&A” (Data & Analytics) are worth unpacking individually.
The first is the most consequential: moving from treating AI as a tweak to treating it as a transformation.
This shift starts and ends with an AI ambition for leveraging AI to transform, not tweak, business and operating models aligned to achieve audacious business objectives.
The second concerns how organisations structure their teams.
Gartner describes the future D&A organisation as one of “tiny teams”: smaller, decision pods of broad-skilled talent augmented by AI and AI agent specialists focused on business outcomes. Pacesetting companies are already experimenting with teams as small as one technical person and one business person. This is a radical departure from how most data functions are currently organised.
The third shift focuses on context as infrastructure.
Organisations with the highest maturity of AI-ready D&A capabilities are achieving up to 65% greater business outcomes, including revenue growth and cost optimisation. D&A success in 2030 is not about better models, it is about giving agents governed, contextual access to the right data.
Context, including semantics and metadata, has become mission-critical infrastructure.
The governance shift is perhaps the most important for organisations trying to move from pilots to production. Only 23% of IT leaders surveyed said they are very confident in their organisations’ ability to manage security and governance when deploying GenAI tools.
Traditional controls should be overhauled to prioritise trust-based governance models for AI agents, without trust in the data, outputs, and decisions of AI models and agents, there is no value from AI.
Value Compounding vs ROI Thinking
Finally, Gartner asks organisations to move beyond ROI thinking entirely: toward what it calls value compounding, where efficiency gains from high-impact investments are intentionally reinvested into growth and innovation rather than banked as cost savings.
The research is a direct challenge to the AI investment thesis most organisations are currently running: spend on the model layer, hope for outcomes.
The data says the causality runs the other way. The organisations extracting real value built the foundation first.
