The AI Adoption Problem

You can't measure AI trust on a 1‑to‑5 scale.

AI adoption fails when leadership guesses what their people are afraid of. Ethnobot doesn't guess. We have real conversations with your team to map exactly where they sit on the AI maturity curve.

You can't survey trust. You can't put a Likert scale on “I'm scared AI is going to take my job.” So we stopped trying.

The numbers

The reality gap, by the data

Synthesised from the largest AI adoption studies of 2023–2026, across 4,600+ organisations in 14 countries.

Deployment
88%
of organisations experimenting with AI right now
McKinsey 2026
Returns
81%
report no meaningful bottom-line gains from AI investment
McKinsey 2026
Project failure
42%
abandoned AI projects entirely in 2025
S&P Global Survey
Scaling wall
85%
of AI initiatives fail to scale beyond the pilot stage
Capgemini 2023
Real usage
17%
daily active users despite ‘70% adoption’ claimed by management
McKinsey 2023
Boston Gov't
0.8%
actual adoption — 200 of 25,000 trained employees actively using AI
Reboot Democracy 2026

“AI has made me feel at times like an expert and a beginner. It's supercharged my learning.”

ExplorerPolicy researcher

“Worry about becoming dependent on AI for critical thinking and losing independent analytical ability.”

ScepticPolicy researcher

Root cause

It's not the technology.
It's the organisation.

Every major research body now agrees: AI adoption failure is cultural, not technical. Companies treat AI as a point solution when it demands full organisational rewiring — across culture, trust, capability, and workflow simultaneously.

The organisations winning with AI aren't the ones with the best tools. They're the ones who understood what their people actually needed before deploying them.

“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
5.3×

higher success rate for organisations that invest in cultural change alongside technology

HBR 2025

4.3×

more likely to sustain financial results when people investment matches technical investment

McKinsey 2026

75%

of all roles need fundamental reshaping for effective AI integration — not just training or tools

McKinsey 2026

The maturity curve

Assist → Augment → Adapt.
Most organisations are one tier lower than they think.

The AAA framework maps exactly where your people operate — not where your AI policy says they should be. Each tier builds on the last. You cannot skip levels. The gap between where leadership believes the organisation is and where it actually operates — that's the Absorption Gap.

01

Assist

Where most teams actually are

AI supports discrete tasks. Staff are in full control.

Using Copilot to summarise meetings, draft emails, or search information. The human leads every process — AI is a productivity add-on, not a partner. Helpful, but it doesn't transform how work gets done or how decisions are made. Most organisations operating here genuinely believe they're one tier higher.

Staff use AI for individual tasks in isolation
No shared workflows or team-level AI practices
Usage is ad hoc, unmonitored, unpolicied
Low trust in AI outputs — heavy manual checking

“Worry about becoming dependent on AI for critical thinking and losing independent analytical ability.”

ScepticPolicy researcher
02

Augment

Where most leaders assume they are

AI becomes a genuine thinking partner. Staff and AI learn from each other.

Bidirectional learning. Workflows start changing — not just individual tasks, but how teams coordinate, decide, and create. Requires robust data foundations, cross-platform integration, and governance frameworks. This is where AI actually starts delivering the ROI organisations were promised. It's also where most rollouts stall because the cultural and infrastructure prerequisites weren't built first.

Team-level AI practices and shared prompting conventions
AI outputs feed back into improved inputs
Governance frameworks governing AI use
Trust is earned through consistent, transparent AI behaviour

“AI has made me feel at times like an expert and a beginner. It's supercharged my learning.”

ExplorerPolicy researcher
03

Adapt

Where AI strategy decks point

AI takes semi-autonomous action with human oversight.

Proactive optimisation. Continuous learning across integrated systems. AI anticipates needs, surfaces risks before they materialise, and executes within defined boundaries. Demands organisational trust at scale, mature governance, and cultural readiness developed over years — not months. Organisations arrive here by compounding the work at Assist and Augment, not by buying a more expensive platform.

AI operates within delegated decision boundaries
Continuous feedback loops between AI and human judgment
Organisation-wide trust and literacy built over time
Governance and ethics infrastructure is robust

“The fine line between hallucination and useful unimagined scenarios.”

ThinkerGovernment reform professional

The seven dimensions of capability maturity

Moving tiers isn't just about tools.
It's about seven interdependent capabilities.

Each dimension must develop in concert. Organisations that invest only in technology while neglecting governance, culture, or inclusion consistently fail at the Augment threshold.

01

Data Foundations

Quality, accessibility, and ethical management of data — the fuel for every tier above Assist.

02

System Integration

Platform connectivity and information sharing. Silos destroy Augment before it starts.

03

Human–AI Interaction

Staff understanding and genuine collaboration capability — not just training, but lived fluency.

04

Governance & Risk

Ethical safeguards and decision protocols. Absent governance is the most common reason Augment stalls.

05

Inclusion & Equity

Equitable implementation across all users. Adoption fails when AI is perceived as a management surveillance tool.

06

Org Readiness

Culture, leadership support, and change capability. The dimension most organisations underinvest in.

07

Business Value

Measurable operational and outcome impact. If you can't measure it, leadership won't sustain investment.

KEY

No dimension is optional

Skipping Governance to move faster always produces the same outcome: cultural resistance that surfaces 6–12 months post-launch when it's far harder to fix.

Read the full AAA research on The Helix Loop →

Original research

The Absorption Gap:
measuring what your organisation actually enables

Most AI assessments measure tool deployment. We measure the delta between what your people can do with AI and what your organisation actually allows them to do. This gap — invisible to dashboards and surveys — is where adoption dies.

Individual capability

What people can do

Given the right conditions, most knowledge workers can use AI to draft, analyse, synthesise, and create at a level organisations have never seen. The ceiling is high and rising.

AI capability ceiling85%
Current individual capability62%
Organisational enablement

What your org actually allows

Policy gaps, trust deficits, missing infrastructure, cultural resistance, and leadership uncertainty suppress what individuals can actually use in practice. This gap is the crisis.

Policy clarity28%
Cultural readiness38%

“I want to see more human-centred AI — a focus on ‘Is this better for stakeholders involved?’ Less focus on efficiency.”

PragmatistActuary

This framework underpins Ethnobot's measurement methodology. 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 →

Watch

The AAA Framework explained in 3 minutes

Watch: The AAA Framework Explained

3 minutes · Suhit Anantula, The Helix Lab

Video coming soon

What you'll receive

Not a score. A real picture of what's happening.

When your team finishes their conversations, you get a diagnostic report — your exact AAA distribution across the six AI archetypes, and the single biggest bottleneck holding your team back. Nothing else.

Your team's report unlocks here

48 hours after conversations complete

View a sample report →
Free diagnostic

Run your free 15-person diagnostic.

Fill this in. Interview links go to your team within 24 hours. No call required.

Or skip straight to the form