Everywhere I look, organisations are rushing to build AI competency frameworks, AI maturity models, and define process best practice for AI return on investment. Beautiful AI-assisted slide decks are being published, stages of maturity are being numbered (pro-tip: it's always 5 levels), and human competencies for the digital era are being classified and rationalised into impressive taxonomies. And then the benchmarking begins…
The hidden assumption is that if we measure AI maturity, then we can manage the process and realise the value. Peter Drucker is often quoted as saying "If you can't measure it, you can't manage it." and this epitomises 20th century thinking. But it is another well known management guru of the 20th century, William Edwards Deming, who prophetically wrote in his 1982 Out of the Crisis that "the most important figures one needs for management are unknown or unknowable."
Frameworks are artefacts of the 20th century, the era of business process management. This was a century that believed progress was linear, rational, and controllable. But 20th century frameworks compress the living dynamics of the digital era into static objects. They trade narratives for neatness, complexity for simplicity, and uncertainty for comfort. And in doing so, they offer the illusion of control, while masking what really drives digital value.
Why Frameworks Fail in a Digital World
The 21st century won't bend to 20th century reason.
Frameworks emerged in an industrial age that valued order, standardisation, and repeatability. It made sense: factories, supply chains, and bureaucracies were systems that thrived on classification and taxonomy guided by specialists. And it really worked!
As someone who built a career in business process management, and having managed a lot of IT over the decades, I understand the deep cultural pull of methodologies and frameworks for colleagues in the many administrative professions that are the bedrock of most organisations.
Performance was constructed on standards, best practices, and compliance regimes. Classification systems, job families, competency models, "best practice" handbooks. Most of the job titles and professions in our organisations were born from the same 20th century faith in order. That faith is hard to let go of. And so begins the search for those same frameworks that fit the new digital realities.
But we must. The world after the connectedness and accompanying complexity explosion of post-globalisation, wealth democratisation and the internet revealed that digital reality is non-linear, emergent, and adaptive. It doesn't behave like a production line.
Patterns and pattern recognition, not frameworks, are the real logic of the digital age. AI itself doesn't "follow" frameworks. It doesn't care about your staged capabilities. It learns from noise, contradiction, and the dense probabilistic universe of pattern recognition. Why should our management of AI be any different?
Digital culture doesn't emerge from frameworks. It emerges from subculture, experimentation, improvisation, and local practice. The technology itself is downstream of the culture, not the other way around. If you don't understand that, you'll be forever surprised by why some teams deliver exponential impact and others stall, despite being in the same "stage" of a maturity model.
The Illusion of Maturity
Let's look at a real example. In one corner of your organisation, a team may be harnessing agentic AI to subvert internal rules in favour of customer delight and co-design of digital products. They're bending policy to produce value through AI.
In another corner, a team is still struggling to clean data, let alone use AI responsibly.
According to a maturity framework, one of these teams is "ahead" and one is "behind." That's the wrong conclusion. The new truth is that neither is more mature than the other.
Culture is hyper-local. AI value emerges differently in different domains, depending on context, pressure, leadership, and micro-culture. What looks like immaturity in one model may be a necessary condition for a different kind of breakthrough in another. The world is no longer linear.
So What Should We Do Instead?
Throw away the maturity model. Start observing, not measuring.
Here's a pragmatic checklist:
- Look for signals of practice. Where are people actually using AI to get work done? Follow the practice, not the plan.
- Assess capacities, not staged capabilities. Can the organisation learn, adapt, and integrate AI into real workflows? That's capacity. Everything else is theatre.
- Interrogate governance reflexes. When a new AI use case appears, does your organisation ask "how do we enable this safely?" or "how do we stop this until it's approved?" Your reflexes reveal your culture.
- Watch for emergent integration. AI value doesn't appear in a single application. It appears when AI starts weaving into the fabric of how work happens, across teams and functions.
It's Less Work
"Where are we on the maturity model?" is a 20th century question. "How can we identify and apply best practice?" is another. The 21st century question is: What vectors matter here, and what patterns are emerging?
Finance has to get serious about unit allocation, dynamic forecasting (not budgeting) and shared accountability. HR has to shift from a rationalising and classifying organisation to a role coach and team or even individual person amplifier. And IT has to stop chasing standardisation and solution delivery and embrace the delivery of capabilities as its only raison d'être.
More work? No. In fact, it's less.
You spend less time creating slides of staged competencies and more time listening to what's happening on the ground. You waste less energy comparing your organisation to artificial benchmarks and more energy cultivating local conditions and talent for adoption and promotion. What you accept instead is uncertainty. But uncertainty isn't risk. Uncertainty is potential. It is the raw material of innovation.
TL;DR
This is not an abstract debate. It's a pragmatic call to action.
If you want AI to deliver value in your organisation, stop measuring maturity and start cultivating culture. Stop compressing emergent practice into static frameworks. Instead, observe, learn, and adapt. Yes, frameworks are comforting. They tell a story of order in a disorderly world. But comfort isn't value. Comfort won't get you competitive advantage. Comfort won't help your people harness AI to change the way your organisation works.
In the digital era, frameworks are the new bureaucracy. They give you the appearance of control while ensuring you fall behind those who embrace emergence.
So, next time someone in your leadership meeting asks, "Where are we on the AI maturity model?" be brave enough to answer:
"We're not. We're observing, learning, and adapting. That's the only maturity that matters."