A new field report finds a widening disconnect between organisations’ expectations of artificial intelligence (AI) and the actual outcomes being delivered by the firms using it.
The global report, published by tech executive and analyst Kristian Kabashi and The Blank Collar, shows that while AI adoption is now widespread across enterprises, most organisations remain at an early stage of capability development.
The State of AI at Work report states that 87% of workers globally use AI at the beginner level and that only an estimated 13% use AI for meaningful work, highlighting a structural imbalance between deployment scale and productive application.
Rather than signalling a lack of access to tools, the findings point to a deeper execution gap.
“Companies have rolled out AI systems across functions, but the majority of employees are still not translating access into consistent, high-value workflows. The result is a growing mismatch between executive expectations of transformation and the incremental, often experimental reality of day-to-day usage,” it says.
The field report was conducted through observational analysis of AI usage patterns in organisations, practitioner and executive insights, as well as behavioural interpretation of how AI is being adopted in real workplaces.
Expectation versus reality
At the centre of the report is a strong warning that organisations are overestimating the immediate impact of AI, while underestimating the behavioural and organisational changes required to unlock value.
The gap is not presented as marginal, but structural, rooted in how work is currently designed and how AI is being introduced into existing systems.
Kabashi explains: “Your company bought AI, but nobody changed. At some level, we’ve seen this with every major technology shift over the years, but in the case of AI, these results are quite disappointing.
“The technology has such vast promise, but, in my view, it’s not being used the right way. Organisations tend to believe that once the tools are deployed, transformation automatically follows, but that is not what we are seeing in practice. There is a gap between availability and meaningful integration, and that gap is what is holding back results.”
The report suggests that many enterprises are treating AI as a “plug-in productivity layer” rather than a capability shift that requires rethinking workflows, decision-making structures and task design.
This has created a situation where deployment is ahead of adoption maturity, and experimentation is not consistently translating into operational value, it points out.
Putting AI to work
A central concept introduced in the report is what Kabashi describes as a “use case desert” − the difficulty employees face in identifying clear, structured applications for AI within their day-to-day responsibilities.
Kabashi expands on this challenge by emphasising that the limitation is not primarily technical skill, but cognitive framing and organisational clarity. “What blocks proficiency is not prompting. People can learn in an afternoon, and they do pick up the tools quickly once they start experimenting. But they get stuck one step earlier, on a blank question, which is: ‘What do I even point this at?’
“They open the tool, summarise one e-mail or draft a short response, and then they stop, because nothing in the actual job comes pre-labelled or structured in a way that tells them what should be delegated to AI. The problem is not access or ability, it is the absence of clearly defined use cases embedded into everyday work.”
According to the report, this lack of defined entry points means AI is often used in isolated, low-impact ways rather than being embedded into workflows that drive measurable business outcomes.
The report argues that this is one of the key reasons why organisations are not seeing proportional returns on investment despite rapid adoption.
Leading the evolution
Beyond individual usage patterns, the report identifies leadership as a critical constraint on AI effectiveness. It argues that senior executives are still developing a clear understanding of where AI delivers tangible value, and how to operationalise it at scale across teams and functions.
Kabashi acknowledges the novelty of the shift while warning against delay in organisational learning and adaptation. “I don’t want to point fingers or suggest that leaders are acting in bad faith.
“This is all so new, and it is not surprising that senior executives are still figuring out what actually works in practice. But there is a narrowing window for learning through experimentation.
“Everyone better get busy pretty soon, because the companies that figure this out − who can rise above 13% meaningful AI use − are going to be strong competitive performers in their industries. Those that delay or treat this as a surface-level tool rollout will find themselves increasingly outpaced by more adaptive organisations.”
The implication is that competitive advantage will increasingly depend not on AI access, but on the depth of organisational alignment and execution capability.
The State of AI at Work concludes that the central challenge facing organisations is no longer whether to adopt AI, but how to convert widespread experimentation into consistent, meaningful capability.
Share
