Evidence Brief — 2026
88% of organisations are experimenting with AI. 81% report no meaningful bottom-line gains. The technology isn't the problem — the organisation is.
McKinsey State of Organizations 2026 · Harvard Business Review · PwC · Capgemini
"The greatest risk is winning the AI frontier but losing the AI era."Eric Schmidt — Former CEO, Google · LinkedIn, February 2026
The numbers
Synthesised from the largest AI adoption studies of 2023–2026, across 4,600+ organisations in 14 countries.
Root cause
Every major research body now agrees: the AI adoption failure is cultural, not technical. Companies are treating AI as a point solution when it demands full organisational rewiring.
"The problem isn't the technology — it's that companies are treating AI as scattered point solutions rather than full organisational rewiring."McKinsey State of Organizations 2026 →
"Cultural resistance represents the dominant barrier while companies allocate only 10% of transformation budgets to change management."Harvard Business Review — November 2025 →
"AI delivers value only when people trust and understand it."Hitachi Executive, via McKinsey research
Cultural resistance data
Case studies
From Boston City Hall to Australia's largest enterprises — the same gap appears everywhere.
Public sector AI plans without change management produce technically capable but culturally unused systems. Adoption rates consistently under 5% at 12-month mark.
AI adoption declining in Q4 2024. Cultural resistance in mining explicitly cited as primary barrier. Frontline workers disconnected from deployment decisions.
Source →Microsoft Copilot users often remain at basic maturity despite heavy usage — the tool enables passive consumption, not active AI partnership.
The solution
McKinsey's 2026 research is unambiguous: winners pursue both halves simultaneously. Most organisations are only doing one.
"The winners will be those who pursue 'double transformation' — technical and organisational simultaneously — reimagining entire workflows from the ground up."McKinsey State of Organizations 2026 →
Tools, infrastructure, integrations, APIs, model deployment. Most organisations do this part.
Culture, trust, capability, workflow redesign, mindset shift. Almost no one does this part — and it's where we work.
higher success rate for organisations that invest in cultural change alongside technology
Harvard Business Review 2025 →more likely to sustain financial results when people investment matches technical investment
McKinsey State of Organizations 2026 →of all roles need fundamental reshaping for effective AI integration — not just training
McKinsey 2026 →Original research
Our original research framework, developed through ethnographic field studies with South Australian organisations. The Absorption Gap measures the delta between what individuals could do with AI and what their organisational environment actually allows them to do.
Given the right tools and freedom, most knowledge workers can use AI to draft, analyse, synthesise, and create at a level their organisations have never seen from them.
Policy gaps, trust deficits, missing infrastructure, cultural resistance, and leadership uncertainty combine to suppress what individuals can actually use in practice.
This framework underpins Ethnobot's measurement methodology and The Helix Lab's AAA maturity model (Assist → Augment → Adapt). Unlike survey-based assessments, we use ethnographic interview methods to capture this gap as it actually exists — not as organisations report it.
Read the research behind this →Our contribution
This evidence synthesis informs our consulting practice at The Helix Lab and the Ethnobot platform. We don't just summarise McKinsey — we conduct primary ethnographic research on why AI adoption fails and how cultural intelligence bridges the gap.
Ongoing field research in South Australian government and enterprise AI implementations using structured qualitative interviews — the methodology surveys can't replicate.
Original IP measuring the delta between individual AI capability and organisational enablement — a metric no existing assessment tool captures.
Anchor clients in local government, manufacturing, and education serve as validation for industry-wide patterns. Each study unlocks a vertical.
A three-stage framework (Assist → Augment → Adapt) that maps individual AI behaviours to organisational transformation readiness — grounded in interview evidence, not self-report.
Most organisations already know they have an adoption gap. Few have the methodology to measure it, or the cultural intelligence to close it. That's where we come in.