The number that should worry every CIO funding an AI initiative
A new Ensono survey found that 61 percent of organizations have paused, delayed, scaled back or abandoned core IT modernization initiatives over the past 24 months, redirecting the money toward AI projects instead. That is not a marginal reallocation. It is a majority of surveyed enterprises deprioritizing the infrastructure and legacy-system work that AI initiatives ultimately depend on, in favor of the AI initiatives themselves, which is a sequencing problem hiding inside what looks like a funding decision.
Ensono's Chief Strategy Officer Brian Klingbeil described the dynamic directly: people are redeploying spend toward AI projects, and there is a lot of pressure to keep up. That pressure is coming from boards and competitors as much as from IT itself, which makes it a harder trend to resist through internal discipline alone. A CIO who correctly diagnoses the trade-off still has to defend a modernization budget against a board asking why the company is not moving faster on AI, and that is a genuinely difficult conversation to win on the data alone, particularly when a competitor's AI pilot is getting press coverage and the modernization work by design produces nothing visible to a board for at least a year.
The 24-month window this data actually covers
It is worth being precise about the timeframe, because it changes how alarming this number should be. Ensono's 61 percent figure covers the past 24 months, which means it describes a sustained two-year pattern of enterprises repeatedly choosing AI spend over modernization spend at the point of budget conflict, rather than a one-quarter panic response to a single AI headline. A sustained two-year pattern is a strategy, whether or not any single organization set out to adopt one deliberately, and it should be treated with the same scrutiny CIOs would apply to any other multi-year capital allocation trend.
That distinction matters for how a board should read this data. A one-time reallocation is a decision that can be revisited at the next budget cycle without much institutional resistance. A two-year pattern has already reshaped how vendor contracts, staffing plans and internal roadmaps are built around AI-first assumptions, which makes reversing course considerably harder even once the modernization backlog becomes impossible to ignore. CIOs inheriting this pattern from a predecessor, or who have been living inside it for two budget cycles already, should assume the reversal will take longer than the twenty-four months the drift itself took to accumulate.
The budgets that survived modernization still overran
The organizations that kept modernization funding intact did not necessarily fare better on cost control. Ensono's data shows over 70 percent of respondents reported that modernization budgets exceeded original estimates, and 27 percent overran by 51 percent or more. That combination, shrinking modernization budgets and the ones that remain running badly over estimate, is a worse position than either problem alone, because it means the case for cutting modernization spend looks financially justified even when the underlying work is more necessary than ever.
This is the pattern that turns a temporary funding pause into permanent technical debt. A modernization project delayed by twelve months to fund an AI pilot does not resume automatically once the pilot proves out. It competes again for the next budget cycle against whatever new AI initiative has emerged in the meantime, and legacy systems that were already straining do not get easier to modernize the longer they are left running. CIOs need to treat a modernization pause as a decision with compounding cost, not a neutral timing choice.
Where the money is actually going instead
WalkMe's data quantifies the shift in the other direction: digital transformation budgets grew from 39.4 million to 54.2 million dollars annually at surveyed organizations, and most of that growth was directed specifically toward AI-related projects rather than a proportional increase across all transformation spending. WalkMe's Global Field CTO KJ Kusch called this a deliberate move, not an accident of budget drift, which matches what boards are actually asking their technology leaders to prioritize right now.
Cognizant's Executive Global Director Michele Grant framed the tension in the most direct terms of any source in this data set: AI budgets are a real competitor for the same money that would otherwise fund modernization. That framing matters because it names the two initiatives as competitors for one pool of capital rather than complementary investments, which is closer to how the money actually behaves inside most enterprises than the aspirational language vendors typically use when pitching both at once.
The uncomfortable data point that undercuts the whole trade-off
Here is the detail that should reframe how CIOs think about this trade-off entirely. SAS research found that only 17.5 percent of organizations have data infrastructure fully optimized for agentic AI readiness. That figure matters because it means the majority of enterprises redirecting money away from modernization and toward AI are funding AI initiatives on top of data foundations that are not actually ready to support them, which is close to the worst version of this trade-off available.
SAS's Stu Bradley, senior vice president for risk, fraud and compliance solutions, put it as plainly as any source in this data set: AI and modernization are increasingly two sides of the same coin. Underinvesting in modernization to fund AI is not really reallocating budget between two competing priorities. In a majority of cases it is quietly starving the foundation the AI initiative needs to work, while still paying full price for the AI layer sitting on top of it.
What the sources agree CIOs should actually do
Across four independently sourced perspectives, Ensono, Cognizant, WalkMe and SAS, the recommendation converges on the same structural fix: stop building separate business cases for AI and modernization, and fund them as one initiative with shared success metrics. A unified business case forces the conversation about data readiness, legacy dependency and infrastructure gaps into the same budget approval where the AI initiative gets funded, rather than leaving modernization to compete separately and lose.
The practical version of this for a 2027 planning cycle is phased funding tied to measurable outcomes across both categories at once, not a modernization line item and an AI line item reviewed in different meetings by different stakeholders. If your organization is planning next year's technology budget as two separate conversations, this data set is a direct argument for merging them into one, with the data-readiness gap SAS identified as the specific line item that should anchor the combined business case.
