Gartner Survey: Only 13% of CIOs Report Significant AI Value — What It Means for ERP Modernization
Gartner's CIO survey found only 13% report significant AI value despite 18-fold spending growth, and only 21% have agentic AI governance. What that means for ERP modernization, with Mead Johnson's 11-month S/4HANA migration as a case study.
Gartner’s annual CIO survey, released at its IT Symposium/Xpo in Orlando, found that only 13% of CIOs have realized significant or transformational value from AI, even though AI spending has risen 18-fold in two years. In the same week, a CIO at Mead Johnson Nutrition described how AI tools helped compress an SAP S/4HANA migration from a projected 17–18 months to 11. Read together, the two stories show where AI is producing measurable value today and why most organizations are still waiting for it.
This article breaks down the Gartner figures, explains the governance gap behind them, and looks at what the Mead Johnson project suggests for ERP modernization.
The Gartner numbers
InformationWeek’s October 19 report from Orlando and Gartner’s keynote press release presented several related surveys. Because they measure different groups and different levels of value, they are best read side by side rather than combined:
| Figure | Survey | What it measures |
|---|---|---|
| 13% | Annual CIO survey, 2,000+ IT executives | CIOs reporting significant or transformational value from AI |
| 18-fold | Gartner analysis | Growth in AI spending over two years |
| 58% | Q2 2026 survey, 1,000+ functional leaders | Leaders seeing some positive value |
| 25% | Q3 2026 survey, 303 CIOs and IT risk leaders | Organizations aligning technology risk with business risk |
| 71% | Q2 survey, 2,000+ CIOs and technology executives | Expect to deploy agentic AI within 12 months |
| 21% | Same survey | Currently have agentic AI governance |
Gartner also noted that use-case success remains about one in five, even as the number of use cases has grown.
The 13% and 58% figures are not contradictory. Many organizations see pockets of value in specific functions; far fewer see value at enterprise scale. Gartner’s Daryl Plummer made the same point, noting that pockets of success are hard to replicate across the enterprise.
The 18-fold spending figure is an industry-wide multiple, not a benchmark for any individual company’s budget.
The CEO view
A separate Gartner survey of 203 CEOs, conducted in April and May 2026 and summarized by StockTitan, found that 44% struggle to demonstrate AI-driven financial returns, while 62% still list revenue growth as a top priority and 75% plan to increase technology investment. Gartner’s David Furlonger argued that the value problem will not be solved by adding more AI to existing ways of working. In other words, investment is set to keep rising even though many leaders cannot yet show returns.
The governance gap
The most actionable finding is the gap between agentic AI plans and governance: 71% of CIOs expect to deploy agentic AI within a year, but only 21% have governance capabilities for it. Gartner’s Kristin Moyer, quoted by InformationWeek, called this “out of alignment,” adding: “It’s not possible to create value faster than AI can create risk in that situation.”
Moyer also explained that aligning technology and business risk does not mean every part of the business gets the same value or carries the same risk; it means having a consistent way to manage that relationship. Only 25% of organizations do so today. A practical sign of misalignment is when an organization’s risk committee and AI steering group use different risk assessments.
Gartner proposed a “Value-Risk Flywheel” framework and applied it to reimagining work, reskilling, reducing AI waste, managing shadow AI, and establishing governance. Of those, shadow AI is the easiest to start on immediately: find out which agents and tools employees are already using. Our article on unsanctioned AI agents covers how.
For organizations in the 71% that are not yet in the 21%, a reasonable rule is that no agentic deployment should go live without a named owner, a defined way to override or stop the agent, and clear limits on the data it can access.
The Mead Johnson ERP migration
CIO.com’s October 20 feature by Grant Gross describes a different kind of AI value. Rama Donepudi, global CIO at Mead Johnson Nutrition, led a 2024 migration from SAP ECC to S/4HANA that was projected to take 17 to 18 months and finished in 11, saving an estimated $5 million to $6 million through the shorter timeline.
Donepudi applied different AI tools to different parts of the program rather than treating AI as a single use case:
| Phase | Tools |
|---|---|
| Development | Generative AI coding copilots to remediate custom programs for S/4HANA compatibility |
| Validation | UiPath AI-enabled bots to compare data between legacy and S/4 environments |
| Training | SAP Enable Now and WalkMe for role-specific training content |
| Communications | GPT tools for program updates |
The cutover went smoothly: the system went down on a Friday evening, and on Monday the business asked whether the switch had happened yet. Donepudi also said that change management became easier because AI produced training and communication materials quickly, and that he wanted IT to use AI for productivity before asking other departments to do so.
Several caveats keep this example in perspective:
- It is one company and one program. The timeline depended on Mead Johnson’s custom code, data, and team.
- The tools were assistive. Copilots and validation bots helped people work faster; they were not autonomous agents making business decisions.
- Large consulting firms still play a role. Donepudi named Deloitte, IBM, and Accenture as the firms CIOs hire for program management and change on large transformations. AI reduced hours inside the program; it did not replace the systems integrator.
- The savings came from time compression, not new revenue.
What analysts say about AI in transformation projects
Gross’s article includes other perspectives. Gartner AI analyst Chandrasekaran said interest is higher among midsize CIOs with limited budgets for multi-year programs and more tolerance for risk, but noted that hallucinations in code mapping and quality assurance remain a concern. “You need to accept or tolerate a certain level of risk to trust your AI to do some things that you were doing yourself,” he said. He recommends putting governance and data infrastructure in place before using AI on the transformation itself.
Vijay Thangella of Avaya claimed that AI-assisted contact-center modernization can reduce months of work rebuilding workflows, routing, integrations, and business rules to a few days. That is a vendor claim and should be evaluated separately. Brian Klingbeil of Ensono said most organizations have not yet used AI to assist digital transformation; Ensono uses it for code discovery and legacy cleanup.
Lessons for ERP modernization
- Use AI for specific, checkable tasks. Code remediation, data validation, and training content are well suited because results can be tested.
- Keep humans accountable. Donepudi took personal responsibility for the program; AI-generated code changes still need review, especially in custom ERP code where a hallucinated function can cause failures.
- Measure value in concrete terms. Months saved and consulting fees avoided are easier to defend than broad productivity claims.
- Put governance first for agents. Assistive tools in a well-run program carry different risks from autonomous agents acting in production systems.
Related coverage
See our analysis of Gartner’s finding that only 4% of CIOs have redesigned roles for AI agents, Gartner’s $2.7 trillion AI spending forecast, IBM’s Client Zero program, and ServiceNow’s AI workflow strategy.
Bottom line
Gartner’s 13% figure shows that enterprise-scale AI value is still rare, and the gap between agentic AI plans (71%) and governance (21%) explains part of the reason. The Mead Johnson migration shows where value is achievable now: targeted, assistive AI inside a well-managed program with clear accountability and measurable time savings. For most CIOs, that is a more realistic model than enterprise-wide transformation through agents.
Sources
- InformationWeek, “Gartner: CIOs must balance AI value with risk” (October 19): https://www.informationweek.com/responsible-ai/gartner-cios-must-balance-ai-value-with-risk-and-fast
- Gartner, IT Symposium/Xpo 2026 keynote press release (October 19): https://www.gartner.com/en/newsroom/press-releases/2026-10-19-gartner-it-symposium-xpo-keynote-release-2026
- CIO.com, Grant Gross, “CIOs look to accelerate digital transformation with AI” (October 20): https://www.cio.com/article/4236122/cios-look-to-accelerate-digital-transformation-with-ai.html
- StockTitan, Gartner CEO survey: https://www.stocktitan.net/news/IT/gartner-says-44-of-ce-os-struggle-to-demonstrate-ai-financial-d33q593611yh.html