The AI Transformation Fitness Test
A snapshot inspection of AI transformation initiatives with senior leaders across 15+ sectors. 3 in 4 are at risk of falling short of their goals.
Most leaders are committed to an AI vision.
Their initiatives are at risk of stalling before they deliver transformation.
In 2026, most AI transformation initiatives have budget and commitment. Few have delivered ROI.
As AI continues to advance, the challenge facing leaders trying to transform their organisation is not a technology one. AI transformation requires change at the level of how people work, not just the tools they have access to.
It is not enough to work faster or rewrite a few more emails a day. The real question is: so what? Does this actually drive company objectives? To achieve ROI from AI, leaders must change how the organisation operates around it. The problem is most leaders are in the dark about both the gap between what their initiative promises and what it delivers, and the operating system required to close it.
At an exclusive Practical AI Summit in partnership with HumanX, we ran the 26-Point GenAI OS Inspection on 64 AI transformation initiatives led by senior leaders (nearly 40% C-Suite or Founder, 60%+ Director-level) across 15+ sectors and 9 countries. The results revealed a gap across every dimension of the operating system for AI transformation.
Most initiatives had the visible starter kit: access (78%), CEO-level vision (66%), a champions network (66%). Most were missing the execution layer that turns commitment into outcomes. 3 in 4 AI transformation initiatives are at risk of falling short of their goals.
For leaders looking to drive AI transformation, this report is a diagnostic of what’s hidden, a picture of what’s possible, and a call to close the gap between intent and impact.
The 26-Point GenAI OS Inspection.
Designed by Jeremy Utley from running AI transformation programmes across Fortune 100 companies, professional sport, enterprise software, and financial services.
The 26-Point GenAI OS Inspection is a structured diagnostic that stress-tests 6 layers of an AI transformation initiative in 26 yes/no questions.
A score below 60% signals an AI transformation initiative at risk of falling short of its goals.
Has leadership articulated a clear, published AI ambition: not just intent but codified commitment?
Has the organisation given people the tools, education, goals, and cultural permission to use AI at work?
Are there formal structures for AI: dedicated leadership, opportunity identification, and a champions network?
Does the organisation track what AI is doing: proficiency, economic impact, performance standards?
Are there clear rules for AI use, and clear permission to experiment with new tools?
Is the organisation avoiding the most common transformation failure modes: compliance-led rollouts, ignoring people, treating AI like standard software?
3 in 4 AI transformation initiatives are at risk of falling short of their goals.
across respondents
60% passing threshold
initiatives at risk
n=64. Senior leaders who ran the inspection on their own AI transformation initiatives at HumanX’s Practical AI Summit.
While the inspection revealed broad commitment to AI, with most having frontier LLM access, a CEO-level vision, and a champions network, few had built what turns commitment into outcomes: ownership, direction, and measurement.
CEOs are committed. Their AI transformation initiatives are missing the mark.
The paradox isn’t about effort. Budgets are approved, visions are published, access is given. Inside every dimension, the commitment layer gets built. The execution layer underneath (ownership, direction, measurement) doesn’t.
AI looks like a technology rollout, so most leaders are managing it like one. The visible parts get built. The parts that deliver don’t.
Enablement 55%. Operating Model 53%. Leadership Vision 52%. Governance 51%. To-Don't 50%. Measurement scored 24.7%.
Five of the dimensions measure whether an initiative is poised to drive impactful transformation. Measurement is a critical check in closing the gap between intent and impact.
The six findings below trace this pattern.
A vision without a manifesto puts the initiative at risk of drift. No one outside the CEO can translate it into strategy or challenge it, and when priorities shift there’s nothing to hold the line against. On several leadership questions, 17–19% of senior leaders could not answer critical questions about their own leadership’s AI commitment. When the senior team can’t describe the vision, the initiative has no anchor.
access
channel
channel
celebrate use
education
proficiency goal
Access without a definition of good use splits the initiative’s outcome. A small group of motivated employees figure out how to use AI well. Everyone else uses it to summarise meetings and rewrite emails. The initiative funds a licence estate with no shared standard for what those licences are for. That’s impossible to measure or scale.
An initiative without named ownership has no one to make the trade-offs or defend the budget when priorities shift. Champions create momentum. A Head of AI keeps the initiative from stalling the first time it hits a hard decision. Without one, the initiative belongs to everyone, which usually means it belongs to no one.
An initiative without measurement can’t course-correct or defend its budget the moment a CFO asks where the ROI is. Leaders end up scaling what isn’t working and missing what is. The initiative doesn’t fail because the idea was wrong. It fails because nobody can see what’s happening.
(what’s permitted, restricted, or requires approval)
(how employees can evaluate and adopt)
When the only clear signal from governance is what’s banned, the initiative inherits a conservative workforce. Employees who aren’t told how to evaluate new tools safely don’t evaluate them at all. The initiative clears compliance. It also suppresses the experimentation culture that produces the compounding returns AI is supposed to deliver.
An initiative rolled out like enterprise software can show activation (logins, training completions, rollout milestones) without producing integration. It looks successful on a deployment dashboard. It doesn’t move the numbers the initiative was set up to move. A SaaS rollout produces users. A teammate onboarding produces collaborators. The framing shapes the outcome.
AI transformation is a human change. Most companies are executing it as a technology one.
Licenses are easy to sign. Visions are easy to publish. Policies are easy to write. Ownership, measurement, and a new way of working are harder. And the harder parts are where AI transformation actually lives. Leaders build what’s visible and assume the rest is happening. It usually isn’t.
"When I survey CEOs, most of them say I've done that. And when I talk to those same CEOs' employees, most of their employees say they haven't done that."
The single largest cluster in the score distribution sits at 50–59%: 17 of 64 respondents. Their scores on Enablement, Governance, and the To-Don't List are all above 60%. Their Measurement score is 23.5%. They have built the parts of an AI transformation initiative that show up on a slide. None can prove it's delivering.
The visible layer is in place. What’s missing underneath has a specific shape: a clear standard, per role, of what good AI use looks like. Without it, nothing can be measured, and nothing can be managed.
"If I had a one-question survey, this would be it: do you have a clearly defined standard of AI proficiency for your own role? Only 23% do. How in the world can you measure it, or manage it?"
The top mandates for AI transformation in 2026.
The gap between commitment and execution won’t close on its own. And the longer leaders wait, the wider it gets. Here’s what the top-performing initiatives did differently to move from AI investment to AI impact.
Among the 10 respondents scoring 75%+. Every single dimension above 76%.
The six mandates below are what they built, and what needs to happen in most organisations to close the gap.
Write the AI vision down, and let it be challenged
A vision that lives in the CEO's head isn't a direction your organisation can execute. Write it down as a manifesto. Make it public, specific, and falsifiable. Companies that do this score 1.5× higher overall.
Define what good AI use looks like in every role
Access to a tool isn't the same as knowing what to do with it. Define a proficiency standard for each role: what should people be able to do with AI that they couldn't before? Fund the education that gets them there. Access without a definition is a licensing cost.
Appoint someone whose job is AI transformation
Champions networks spread enthusiasm. They don't create accountability. A named executive owner (a Head of AI or equivalent) is the difference between activity and outcomes. Companies with one score 1.5× higher overall and 2.1× higher on Measurement.
Build measurement before scaling further
Measurement is the single strongest predictor of overall score in this inspection. Start simple: a proficiency standard for each role, a way to track economic impact, AI criteria in reviews and hiring. Without these, you are guessing.
Match restrictions with permissions
Most organisations have told people what they can't do with AI. Fewer have shown them what they can. Publish a clear policy for trying new tools. Tell people how to evaluate, where to flag, how to escalate. Enabling experimentation builds culture. Restricting it doesn't.
Roll out AI like onboarding a teammate, not deploying software
A SaaS rollout optimises for activation. A teammate onboarding optimises for integration. Organisations that treat AI like a new teammate score 1.7× higher overall and 3× higher on Measurement. The framing shapes the outcome.
Mindstone is an AI transformation company that helps leaders turn AI commitment into company-wide practice. We combine strategy, structured education, champion infrastructure, and an agentic AI platform to build the operating system your organisation needs to work at a different level.
Here’s what we can help with:
We work with your leadership team to build what turns commitment into execution: named ownership, a defined transformation plan, measurement systems, and the change management required to lead a human transformation, not just a software one.
Role-specific proficiency standards and the sustained education that gets every employee there. From AI foundations to agentic capability: not a one-time training event, but a programme that builds real fluency across the whole organisation.
Named AI champions, written standards, and the platform infrastructure that keeps momentum going after the initial push. This is what makes the difference between a pilot that fades and a transformation that compounds.
We embed Rebel across your organisation: an AI platform that compounds with every meeting, decision, and workflow. Shared memory, team-wide automations, and 90+ tools connected to how your organisation actually works.
Mindstone helps organisations build the operational system for AI transformation. The 26-Point GenAI OS Inspection is the diagnostic we use to assess where that system exists and where it doesn't. This report is what senior leaders at HumanX 2026 discovered when they ran it on their own AI transformation initiatives. All scores are self-reported. All data is anonymous.
Where does your AI transformation initiative stand?
Schedule a 30-minute consultation to accelerate the impact of your AI transformation initiative.
Book a meeting →