Most AI projects stall at pilot stage, survey finds

Orgvue found that 78% of organisations have seen AI projects fail or stall at the pilot stage, according to its annual survey of more than 1,100 senior decision-makers at medium and large organisations.

The research highlights a gap between investment and execution. While 92% of organisations have invested in AI, 35% reported project failures and 43% said projects were still stuck in pilot.

Another 32% said they still did not understand how to make AI work, while 84% said their organisation should have a deployment roadmap with clear return-on-investment targets. The figures suggest many companies are under pressure to justify spending even as they struggle to turn projects into operational results.

Competitive pressure appears to be a major factor in decision-making. Some 57% of business leaders said they deployed AI because competitors had done so, and that projects stalled or failed because deployments were rushed.

Investment rises

Spending plans continue to rise despite those setbacks. Orgvue found that 83% of organisations plan to increase AI investment this year, up from 80% a year earlier and 77% in the previous survey.

The proportion of organisations that increased AI investment by 50% or more rose to 31% from 27% a year earlier. The share planning an increase of 50% or more this year also climbed, to 35% from 31%.

Higher spending has been accompanied by a stronger focus on oversight and measurable outcomes. Concerns about employees using AI without proper supervision remained the most common fear, cited by 45% of respondents, little changed from 47% a year earlier.

At the same time, 28% said they feared choosing the wrong solution and wasting investment, up from 23% the previous year. A further 23% said they feared being left behind if their organisation did not improve its understanding of the technology.

Workforce focus

The study suggests companies increasingly see workforce planning as central to AI deployment. Nearly half, 49%, said they were reskilling employees to prepare for AI, while 44% said they had increased learning and development budgets to ensure staff received the right training.

Policy work is also expanding. Some 52% said they had introduced policies governing how AI is used in the workplace, up from 46% two years earlier.

Even so, organisational readiness remains a problem. Orgvue found that 65% of organisations expect AI to drive workforce change in the next 12 months, but 34% said they lacked the expertise to manage that change and 26% said structural issues were holding back readiness.

The data also suggests confidence remains relatively high despite operational difficulties. Some 73% of respondents said they expected to be taking full advantage of AI by the end of the year, up from 69% in the prior survey.

That confidence sits alongside signs of persistent weakness in planning. In the previous year's study, 27% said they did not have a clearly defined AI roadmap and 25% said they did not understand which roles and jobs would benefit from the technology.

Oliver Shaw, chief executive of Orgvue, said the latest findings show a shift in how business leaders are approaching the technology.

"2024 was the year of optimism for AI, while in 2025 we saw more caution from businesses that learnt the hard way that AI deployment can go badly wrong. In 2026, we see an urgency from business leaders to begin delivering value and to reshape the workforce before their competitors do.

"Rising investment shows that organisations still believe in the transformative power of AI, but few businesses are taking the time to understand their workforce structure as it is today. Failed AI deployments are not a technology problem; they're a workforce design problem," Shaw said.

The survey covered 1,163 senior decision-makers, including board executives, investors, C-suite leaders and senior managers. Respondents came from organisations with at least 2,000 employees in the United States, Canada, the United Kingdom and Ireland, and at least 500 employees in Australia, Hong Kong, Malaysia and Singapore.

Shaw said many businesses were still moving too quickly into automation without enough clarity about jobs and skills.

"Too many organisations are rushing into workforce automation before they understand what they're trying to automate. They haven't thought through how new technology and processes will change work, the roles needed to do that work, and the skills needed to perform it. Smart leaders understand their current roles, skills, costs and capacity before they make decisions like this.

"The stakes are high, but only workforce design can turn AI into an engine for growth. Ninety percent of business leaders told us they're deploying AI as part of their growth strategy, yet a third admit they still don't know how to make it work for their organisation.

"AI will not deliver long-term growth on its own. But when deployed intentionally, grounded in workforce data, role clarity and human capability, it becomes one of the most powerful tools leaders have to shape the organisation's future," he said.