We talk a great deal about the speed of AI implementation. We rarely talk about the cost of adoption. That imbalance is where most AI value goes to die, and the data is unambiguous about why.
There is a familiar pattern emerging in organisations that have moved quickly to deploy AI. The tools are live. The dashboards are built. The vendor has presented the ROI case. And yet, six months later, the technology is being used partially, inconsistently, or not at all by the people it was designed to support.
This is not a technology failure. It is an adoption failure. And it is happening at scale.
In a May 2025 survey of 506 CIOs and technology leaders, Gartner found that 72% reported their organisations were breaking even or losing money on their AI investments. Separately, the McKinsey Global AI Survey found that while 88% of organisations are now using AI in at least one business function, only around 30% of AI pilots successfully transition to scaled impact.1,2
The technology is advancing faster than organisations are preparing their people to use it. That gap, between deployment and genuine adoption, is where transformation stalls and investment fails to return.
The Hidden Cost That Most Leaders Are Not Calculating
The numbers that frame the scale of AI implementation are well known. The numbers that frame the cost of making it stick are much less discussed.
Gartner research is direct on this point: for every 100 days an organisation spends implementing an AI tool, it should expect to spend an additional 25 days training staff, and a further 100 to 200 days on change management. That is up to 300 days of human investment sitting behind every 100 days of technical deployment.3
72%
of CIOs report their organisations are breaking even or losing money on AI investments. Gartner, May 2025, survey of 506 CIOs and technology leaders
Most business cases for AI investment model the technology cost in detail. Very few model the change management cost with the same rigour. The result is a consistent and predictable shortfall: organisations that budget for the algorithm and underfund the adoption find themselves owning an expensive capability that their teams are not equipped, confident, or behaviourally ready to use.
Gartner's description of this dynamic is precise: every AI tool an organisation buys comes with ten hidden costs that were not anticipated. The transition mortgage, as Gartner's analysts put it, keeps being paid long after go-live, in the form of retraining, workarounds, and the slow erosion of confidence in a system that was never truly embedded.
Why AI Fails: The Pattern Is Consistent
The research on why AI transformation fails points consistently to the same causes. They are not technical. They are human.
A 2025 peer-reviewed study published in the International Journal of Human-Computer Interaction concluded that AI adoption initiatives fail at an alarming rate, and that the primary reason is not the AI itself, but the absence of a behavioural approach to change. Many organisations implementing AI focus predominantly on the systems, assuming people will fall in line. They do not.4
The failure modes cluster around four recurring patterns. People are not prepared for the shift in how they work, not just what tools they use. Processes are not redesigned around the new capability: the AI is layered onto existing workflows rather than integrated into new ones. Leaders underestimate the behavioural change required, treating adoption as a training event rather than an ongoing transformation. And organisations approach AI as a plug-and-play upgrade, when in reality it requires people to fundamentally change how they think and make decisions.
"AI adoption is not a technology problem. It is a behavioural one. Users must learn not just new interfaces, but new ways of thinking about work itself, and that is not a challenge a workshop can solve."
1 in 50
AI initiatives achieve true transformation, according to Gartner analysis. The odds of simply achieving ROI are one in five. Gartner Symposium 2025
The odds cited by Gartner, one in five for ROI and one in fifty for true transformation, should give every leadership team pause. They are not a counsel of despair. They are a clear signal about where the investment needs to go.
The Supply Chain Context: Where We See This Every Day
In supply chain planning, where AI is being deployed at pace across demand forecasting, inventory optimisation, and S&OP processes, the adoption challenge takes a specific and well-documented form.
Planning teams are asked to work with systems that generate recommendations algorithmically, recommendations that require interpretation, judgement, and the confidence to act on or override. That is a meaningfully different kind of work from the planning many of these teams have done for years. It requires new skills, yes. But more fundamentally, it requires new behaviours: a different relationship with the data, a different posture in the planning meeting, a different understanding of where human judgement adds value and where the system should be trusted.
When those behaviours are not developed, when training stops at system literacy and does not extend to decision confidence, process ownership, and genuine capability, the pattern is predictable. Planners revert to spreadsheets. The AI recommendation is noted and ignored. The forecast accuracy improvements promised in the business case do not materialise. And within two years, the organisation is considering whether the system was the right choice.
The system was not the problem. The adoption was.
What Getting This Right Actually Requires
Organisations that successfully scale AI, the minority that McKinsey identifies as high performers, share a set of characteristics that are worth examining carefully.
They commit a disproportionate share of their AI investment to people and processes, not just technology. Research cited in the California Management Review found that organisations investing 70% of their AI resources in people and processes, rather than in the technology itself, achieve ROI significantly above their peers.5
They have senior leaders who actively champion AI adoption, not just authorise it. McKinsey's 2025 research found that high performers are three times more likely than their peers to have senior leaders who demonstrate visible ownership and commitment to AI initiatives, who role model the use of the technology and actively drive adoption rather than delegating it.
And they treat adoption as a continuous process, not a launch event. The nature of AI is that the tools keep changing, which means the training requirements keep changing too. Organisations that build ongoing capability development into their operating model, rather than delivering a one-time programme at go-live, are the ones that sustain the gains.
30%
of AI pilots successfully transition to scaled impact across the enterprise. McKinsey Global AI Survey, 2025
The implication is clear. Investing in AI without investing equally in the human system around it is not a partial success. It is a predictable route to the 70% of transformation programmes that fail to deliver their stated objectives.
The Grange Perspective: The Human System Is Where Value Is Won or Lost
At Grange, we work with supply chain planning teams every day who are navigating exactly this challenge. We see the gap between AI deployment and AI adoption up close, and we understand what it costs when that gap is not addressed.
Our conviction is that digital planning, including AI-enabled planning, only delivers value when people adopt it. That means investing just as heavily in skills, confidence, behaviours, and change leadership as in the algorithms themselves. It means pairing deep planning expertise with human-centred change enablement, from the earliest stages of programme design.
AI is no longer a technology project. It is an organisational transformation. The companies who understand this, and invest accordingly, will move faster, achieve better adoption, and outperform their markets.
Is your AI investment matched by the right investment in your people?
Grange helps organisations close the gap between AI deployment and AI adoption, through supply chain planning expertise, human-centred change enablement, and capability building that makes new ways of working stick.
Book a conversation with the Grange team →
References
1. Gartner (2025). Survey of 506 CIOs and technology leaders on AI investment returns, May 2025. Cited in: Gartner IT Symposium/Xpo, October 2025.
2. McKinsey & Company (2025). The State of AI: Global Survey. July-August 2025, 1,993 respondents across 105 nations.
3. Gartner (2025). Organizational Entrenchment: Is Your Organization Wired to Resist AI? Gartner IT Symposium/Xpo keynote data, October 2025. Presented by Daryl Plummer, Gartner VP Analyst.
4. McGuire, J. & De Cremer, D. (2025). Leading AI Adoption in Organizations: Introducing a Behavioral Human-Centered Approach. International Journal of Human-Computer Interaction. Published online 18 July 2025.
5. Chopra, A. (2025). Bridging the Gaps in AI Transformation: An Evidence-Based Framework for Scalable Adoption. California Management Review Insights, November 2025.
Statistics cited are drawn from published industry research and peer-reviewed academic sources. Readers are encouraged to verify figures against the latest available editions of each report, as data is periodically updated.