Everyone's Piloting AI. Almost Nobody Has Shipped It.

Only 11% of organisations have scaled AI agents across the enterprise. That is KPMG's number, and it sits next to another one: 95% have a formal AI strategy and plan to spend a weighted average of $186 million on AI in the next year.

Usage is not value

McKinsey finds 88% of organisations regularly use AI in at least one function. Only 44% say it is scaling across the enterprise, and only 37% report any EBIT impact. The share of "AI high performers", those attributing 5% or more of EBIT to AI, has stayed at 6% even as usage has climbed. Deloitte found just 25% have moved 40% or more of their pilots into production. MIT NANDA's headline is the starkest: 95% of GenAI pilots show zero P&L impact.

It is the organisation, not the model

MIT calls it the "learning gap": pilots cannot retain feedback, adapt to context or improve over time, while employees quietly use personal AI tools anyway. Deloitte describes "pilot fatigue". Pilots are cheap, visible and politically easy, so governance, integration and workflow redesign get pushed back. Size does not fix it. 40% of $1 billion-plus organisations are scaling agents somewhere (McKinsey), yet KPMG's sample, mostly $1 billion-plus, shows only 11% at true enterprise scale. Only 8% say they can measure the return.

This is the J-Curve

Performance dips before it improves, because the organisation adopted a new capability faster than it redesigned the work, governance and incentives around it. Most of the survey data is a snapshot of thousands of companies sitting in that trough at once.

What the leaders do

Deloitte's "superscalers", who have scaled 70% or more of their pilots, give as much of the workforce as possible access to AI, positioned to support what people already do well. The practical lesson from four surveys: by late 2026 the constraint is rarely the model. It is almost always the organisation.

A caveat: each survey defines "pilot", "production" and "scaling" differently, so read them as convergent evidence, not one comparable statistic.

The research behind it: The State of AI Adoption, 2026 Edition

David Viney built WPP Open, a $350m-a-year agentic AI platform, and now runs Alchemy Consulting from London. He works as a fractional CIO in London, a technology board advisor for founders, CEOs and chairs, and an AI governance and strategy consultant; built on his MASTER-AI framework. He leads enterprise ERP transformation programmes, provides cyber security consulting in London - including fractional CISO cover - and offers technology leadership coaching for CIOs and heads of IT. Most engagements begin with a fixed-price Technology MOT.

Posted using PALnet