The technology was never the hard part
A new benchmark survey of senior data and AI executives across nearly 110 Fortune 1000 and global organizations found a lopsided answer to a deceptively simple question: what actually blocks data transformation from succeeding. 93.2% of respondents pointed to cultural challenges and change management, the hard work of getting people across the organization to define and use data differently in how they actually make decisions day to day. Only 6.8% blamed technology limitations of any kind. For a function that has absorbed enormous capital spend over the past decade on platforms, pipelines, and now a fresh wave of AI tooling on top of all of it, that lopsided ratio should force a genuine reset in how most data transformation budgets get allocated going forward.
Petar Bielovich, General Manager of Data & Analytics at Atturra, who published the analysis of the survey, draws a sharp distinction worth adopting as a working definition inside your own organization and repeating in every steering committee meeting: "A data platform is delivered when the technology works. A data transformation is realised when people repeatedly work and make decisions differently because of it." Most programs quietly report success against the first, easier definition at every quarterly review, while failing the second, harder one without anyone on the steering committee formally measuring it or raising the discrepancy out loud.
Gartner's number confirms the direction, not just the survey
This is not an isolated data point. Gartner research cited in the analysis found that 61% of organizations are already changing their data and analytics operating model specifically because of AI, meaning the org chart, decision rights, and workflow around data are moving, independent of any single vendor's platform choice. That is a meaningfully different signal than another survey about AI adoption intent, because operating model change is expensive, slow, and usually only happens when leadership has concluded the current structure cannot deliver.
Put the two numbers together and the implication is specific: the technology procurement decision CIOs spend most of their review cycles on is a smaller lever than the organizational design decision that usually gets delegated to HR or left unscoped entirely. If 61% of peer organizations are already restructuring around this, a data transformation program that only budgets for platform licensing and implementation services is behind the market, not ahead of it.
Designing for decisions, not dashboards
The analysis argues that most data transformation programs are scoped backward from the start, beginning with the technology rollout itself and then simply hoping that behavior change follows naturally once the new dashboards and pipelines are live and training sessions have been scheduled. The alternative it proposes is to start deliberately from the specific business decisions an organization most wants to improve, demand forecasting accuracy, pricing response time, churn intervention timing, and then design the entire data and AI investment around measurably changing how those particular decisions get made by the people responsible for them.
This reframing has a practical budgeting consequence. A technology-first program measures success in uptime, data quality scores, and user logins, all of which can look healthy while nothing about actual decision-making has changed. A decision-first program measures success in whether a specific decision is being made with different inputs, by different people, on a different timeline than before the investment. That second measurement is harder to report in a quarterly business review, which is exactly why most programs avoid it and default to the easier technology metrics instead.
Where the money actually needs to go
If 93.2% of executives say culture and change management are the dominant barrier, the obvious next question is whether transformation budgets reflect that ratio. In most organizations, they do not. Change management is still frequently funded as a thin layer of training and communications bolted onto a technology-heavy program, rather than resourced as a core workstream with its own budget, timeline, and executive sponsor independent of the platform vendor's implementation team.
The fix is not complicated to describe, even if it is organizationally uncomfortable to execute: fund behavior change at a ratio closer to the actual barrier it represents. That means dedicated roles focused on how specific teams will use data differently day to day, success metrics tied to decision behavior rather than system adoption, and a willingness to slow the technology rollout if the organization is not ready to change how it works around it.
The decision this survey is really asking CIOs to make
Every CIO currently sponsoring a data or AI transformation program should take this survey as a direct prompt to audit their own budget split between technology spend and change management spend against the 93.2 to 6.8 ratio executives report as the real source of program failure across the industry. If your program's current spend still mirrors the older assumption that technology implementation is the hard part of the work, you are systematically funding the smaller 6.8% problem while leaving the much larger 93.2% problem under-resourced and, in most organizations, without a single named owner.
This is also a governance question that extends well past simple budgeting arithmetic. A transformation scoped around specific decisions, with named owners who are accountable for whether those decisions actually change in practice, is far easier to audit and defend credibly to a board than one scoped around generic platform milestones like uptime or user counts. Given that 61% of peer organizations are already restructuring their operating models around exactly this reality, treating change management as a minor line item rather than the core deliverable of the program is no longer a defensible position for any sponsor to carry into next year's budget cycle and present with a straight face.
