Insights

AI Project Failure Statistics - everything you need in 2026

A data-led view of why AI projects fail, where pilots stall, and what separates high-performing AI programmes from activity without outcomes.

Riaz Kanani ai ai adoption business strategy decision making

The number that should concern every CEO making AI bets right now: 88% of organisations now use AI in at least one business function. Only 39% see any measurable impact on their bottom line. (McKinsey, 2025)

That gap is not an AI problem. It is a business problem.

The technology is working. What is not working is the way most organisations decide what to do with it, define whether it succeeded, and build the foundations that AI actually needs to deliver results.

Here is what the data tells us about why AI projects fail — and what separates the organisations getting real results from the ones generating activity without outcomes.

We update this page as new research becomes available.

Key statistics at a glance

  • 80%+ of AI projects fail to deliver their intended business value (RAND Corporation, 2025)
  • 88% of organisations use AI. Only 39% see EBIT impact (McKinsey, 2025)
  • $547 billion of $684 billion invested in AI in 2025 failed to deliver intended results (RAND, 2025)
  • Only 6% of organisations qualify as AI high performers seeing meaningful financial returns (McKinsey, 2025)
  • 84% of AI project failures are caused by leadership and process issues, not technology (VentureBeat, 2024)
  • 42% of companies abandoned most of their AI initiatives in 2025, up from 17% in 2024 (S&P Global, 2025)

Overall AI project failure rates

  1. Between 80% and 95% of enterprise AI projects fail to deliver their intended business value — depending on how failure is defined (RAND Corporation, 2025; ResearchGate, 2024)
  2. Of the $684 billion invested globally in AI initiatives in 2025, over $547 billion failed to deliver intended results — an 80%+ failure rate on capital deployed (RAND Corporation, 2025)
  3. 88% of organisations now use AI in at least one business function. Only 39% report any measurable EBIT impact (McKinsey Global AI Survey, 2025)
  4. Nearly 80% of companies using generative AI report no significant bottom-line impact (McKinsey, 2025)
  5. Only 6% of organisations qualify as AI high performers — those seeing meaningful, sustained financial returns from their AI investments (McKinsey, 2025)
  6. Only 1% of organisations describe their AI rollout as mature, with embedded AI across multiple business functions (McKinsey, 2025)
  7. 42% of companies abandoned most of their AI initiatives in 2025, up sharply from 17% in 2024 (S&P Global Market Intelligence, 2025)
  8. Only 48% of AI projects that enter the development process make it all the way to production deployment (S&P Global, 2025)
  9. The average organisation scrapped 46% of its AI proofs-of-concept before they reached production (S&P Global, 2025)
  10. Only 5% of AI pilot programmes achieve rapid revenue acceleration (MIT NANDA Initiative, 2025)

Where AI projects stall: the pilot problem

The failure often does not happen in production. It happens in the transition from pilot to operational deployment — and for a consistent set of reasons.

  1. 30% of generative AI projects are predicted to be abandoned after proof of concept (Gartner, 2024)
  2. Over 40% of agentic AI projects are predicted to be cancelled by the end of 2027 (Gartner, 2024)
  3. 60% of AI projects that lack AI-ready data will be abandoned before delivering value (Gartner, 2025)
  4. The average organisation scrapped 46% of AI proofs-of-concept in 2025 before they reached production (S&P Global, 2025)

Pilots are designed to succeed. They run on cleaned data, with engaged people, in controlled conditions. Scaling requires none of those conditions to hold. This is the pilot-to-production gap where most AI value is lost — and it is almost always caused by problems that precede the technology.

Why AI projects fail: root cause data

The headline many organisations reach for — technology problems — is not what the research supports. The causes are almost always upstream of the tools.

  1. 84% of AI project failures are attributed to leadership and organisational issues, not technology failures (VentureBeat, 2024)
  2. 73% of failed AI projects lack clear executive alignment on what success looks like before the project starts (McKinsey, 2025)
  3. 68% of failed projects underinvest in data governance and foundational systems (McKinsey, 2025)
  4. 61% of organisations that fail at AI treat it as an IT project rather than a business transformation (McKinsey, 2025)
  5. 85% of AI projects fail due to poor data quality or lack of relevant data (Gartner, 2025)
  6. When AI projects are abandoned, the stated reasons are: data quality issues insurmountable (38%), business case no longer viable (29%), loss of executive sponsorship (21%), technical approach infeasible (12%) (Gartner, 2025)
  7. Companies that see significant financial returns from AI are 2x more likely to have redesigned their end-to-end workflows before selecting technology or modelling techniques (McKinsey, 2025)
  8. Organisations where senior leaders actively champion AI adoption perform significantly better than those that delegate AI ownership to a technical team (McKinsey, 2025)

The technology is almost never the cause of failure. The most common culprit is something that precedes the technology entirely: an unclear definition of what the project is supposed to change, and what evidence would indicate it had worked.

The gap between AI activity and AI outcomes

The organisations spending the most on AI are not the ones seeing the best returns. The correlation between AI investment and AI results is weak. What matters is how projects are set up, not how much is spent.

  1. 88% of organisations report using AI in at least one business function (McKinsey, 2025)
  2. Only 39% report measurable EBIT impact from their AI investments (McKinsey, 2025)
  3. 80% of companies using generative AI report no significant effect on their bottom line (McKinsey, 2025)
  4. Organisations with the highest AI spending are not consistently the organisations seeing the best returns (McKinsey, 2025)
  5. 42% of companies abandoned most AI initiatives in 2025 — a 147% increase from the prior year (S&P Global, 2025)
  6. Despite $301 billion in global AI spending in 2026, only a small fraction of organisations report financial returns commensurate with investment (IDC, 2026; McKinsey, 2025)

What the top 6% do differently

McKinsey’s research consistently identifies a small set of behaviours that distinguish high-performing AI organisations from the majority. They are not particularly complex. But they are consistently applied.

  1. High performers define a specific, measurable success signal before any project starts — not “improve productivity” but a concrete behaviour that changes (McKinsey, 2025)
  2. High performers redesign the workflow the AI will operate in before deploying the AI, not after (McKinsey, 2025)
  3. High performers maintain active senior sponsorship throughout the project, not just at kickoff (McKinsey, 2025)
  4. High performers treat adoption as the primary implementation challenge and invest in change management at the same level as technology (IBM, 2026)
  5. High performers run fewer AI projects simultaneously, learn from each one, and build from demonstrated results (McKinsey, 2025)
  6. High performers are 2.5x more likely to have a documented process for deciding which AI projects to pursue (McKinsey, 2025)
  7. High performers pilot inside real processes rather than alongside them — so what they test is what they will scale (McKinsey, 2025)

The adoption problem inside the failure problem

Even projects that reach production often fail to deliver because the people who should use the output do not. This is the part of AI failure that is hardest to see on a dashboard.

  1. 85% of employees have the capability to use AI tools. Only 25% use them regularly — a 61-point gap (IBM Global CEO Study, 2026)
  2. 83% of CEOs say AI success depends more on people’s adoption of the technology than on the technology itself (IBM, 2026)
  3. 54% of C-suite executives say adopting AI is tearing their company apart (Writer, 2026)
  4. 79% of organisations face significant AI adoption challenges — a double-digit increase from the prior year (Writer, 2026)
  5. 39% of organisations lack any formal plan to drive business value from the AI tools they have already deployed (Writer, 2026)
  6. 48% of executives describe their organisation’s AI adoption as a massive disappointment (Writer, 2026)
  7. Team resistance to AI tools ranks among the top three barriers to AI success across all company sizes (IBM, 2026)
  8. Between 2026 and 2028, organisations expect 29% of employees to require reskilling for a different role and 53% to need upskilling to perform their current role more effectively (IBM, 2026)

Final thought

The data makes the same point in many different ways. Most AI projects do not fail because AI does not work. They fail because success was never defined, the process was not ready for it, and the people who needed to change how they worked were not brought along.

The organisations getting the most from AI right now are not the ones with the biggest budgets or the most tools. They are the ones who asked “what will change and how will we know?” before they asked “which tool should we use?”

For a practical framework on setting up AI projects to actually deliver, see Most AI Projects Don’t Fail Because of the Technology. And for a structured way to assess where your business stands before investing further, the AI Readiness Assessment is a useful starting point.

Sources include: McKinsey Global AI Survey (2025), RAND Corporation (2025), S&P Global Market Intelligence (2025), Gartner (2024, 2025), IBM Global CEO Study (2026), Writer Enterprise AI Adoption Report (2026), MIT NANDA Initiative (2025), VentureBeat (2024), IDC (2026). This page is updated continuously as new research becomes available.