The true measure of artificial intelligence's transformative power within any organisation lies not merely in its deployment, but in its effective adoption by employees. Yet, this critical phase of integration presents a significant hurdle. Recent Gartner research indicates that a substantial 72% of employees report difficulty incorporating AI tools into their daily routines. Despite 97% of CEOs worldwide planning extensive AI integration, organisations still face widespread challenges in transforming work practices highlighting a clear disconnect between strategic ambition and the realities of knowledge-worker adoption (Cisco, February 2025).
This gap is further compounded by a sobering acknowledgment from the top: the same Cisco study revealed that a mere 1.7% of these CEOs feel their organisations are fully prepared for this monumental AI shift. Such a profound lack of organisational readiness inevitably amplifies the difficulties faced by employees, placing an even greater onus on individuals to navigate new systems and processes often without adequate overarching support. Addressing this confluence of high ambition, low preparedness, and direct employee adoption barriers is therefore paramount for any organisation aiming to genuinely capitalise on AI to increase productivity.
Reports from the individual contributors themselves tell a similar tale. Around a year ago, Slack found that while 76% of workers felt a sense of urgency to become "AI experts," a mere 33% were even using AI in their daily grind. They also found that these very people spent about 41% of their time doing things that are “low value, repetitive or lack meaningful contribution to their core job functions.” Fast-forward just over a year, and a recent Google report found that only 34% of knowledge workers use AI regularly at work. On the surface, these numbers, when put together, seem ridiculously at odds with each other. Where is the adoption?
This adoption challenge is compounded when we look at the track record of more complex AI endeavours. For instance, looking at enterprise AI pilots, on average, only 5 out of 37 actually make it into production, and a mere 3 of those deliver tangible business impact. Looking ahead, Gartner predicts that in 2025 at least 30% of generative AI projects will be abandoned during or right after the proof-of-concept stage. The reasons cited often boil down to persistent issues like poor data quality, inadequate risk controls, escalating costs, or an unclear business value.
So, while many organisations are investing time and resources into developing bespoke, in-house AI projects, it's worth remembering that a vast array of powerful AI tools are already all around us, readily available, and often just waiting to be leveraged. Loads of them are free to use, at least to a degree. You don’t have to learn how to vibe-code with Codename Goose and integrate a bunch of MCP servers to be productive (though it’s worth a shot). Most knowledge workers aren’t even doing the basics like using Custom GPTs for repetitive tasks, or Deep Research mode embedded in many LLMs for quickly understanding a complex topic.
The truth is, access to the technology itself is the easy part. We are all too often resistant to change, and change is evolving faster than ever. The cracks are starting to show.
So, with these powerful tools readily available and offering such significant promise to boost productivity, why are we leaving them untapped? The answer lies in a series of behavioural barriers, predominantly:
Fear and mistrust
Inadequate change management
Cultural misalignment
Lacking incentive to learn
Balancing learning new skills with existing commitments
Most of these fall squarely back on the shoulders of the employer or organisation. Let’s start with fear and mistrust. More than a few knowledge workers still harbour anxieties about AI's impact on their jobs. For example, Deloitte recently found that around half of workers expect to have to change jobs due to AI threatening their employment. There's also the fear of looking incompetent or being seen as cheating by using AI. While there is a sense that we are now moving beyond that initial fear, half of individual contributors still feel uncomfortable telling their manager that they used AI for work tasks. This lack of trust and psychological safety can lead to silent resistance, with staff quietly avoiding or even unintentionally sabotaging new AI tools, regardless of their potential benefits.
Organisational culture can be change’s greatest ally or its staunchest foe. Risk-averse, siloed, or change-resistant cultures will naturally view AI with suspicion. A striking 71% of organisations point to cultural resistance as a major barrier to AI adoption. This is often compounded by poor, fundamental change management. Gartner’s 2018 paper on cultural tensions says it all - 80% of employees experience cultural tensions or competing priorities they don’t know how to balance when change is introduced.
There seems to be a large gap between leadership's perception and employees' reality: 74% of leaders believe they involve employees in (generic) change initiatives, yet only 42% of employees feel included. Furthermore, a recent Kearney and Futurum study revealed a significant disparity in AI confidence levels between CEOs and managers. 78% of CEOs are highly confident in their leadership regarding AI, but only 28% of mid-level managers feel similarly about their company's preparedness. This breeds an impression of mistrust and a sense that AI projects are being done to them, not with them.
Organisations also need to consider change management in a more traditional sense, and it’s evident that with the ride of GenAI, our ability to manage change has gone the other way. Coinciding with this timeline, in 2023, WTW found that 43% of employees viewed their organisation as effective at managing change, which is a 17 point drop from the 60% mark that it was in 2019, prior to genAI being readily available. Traditional, lightweight change management frameworks such as Kotter's 8 Steps for Leading Change are perfect for this kind of thing, particularly because for many companies right now, leaning into AI is becoming urgent and there are significant risks to late adoption.
Failing to clearly communicate the intent and benefit or perceived value behind an AI rollout leaves a vacuum filled by fear of the unknown and a lack of understanding of why the change is happening, resulting in cultural misalignment. 2019 Gartner research highlights a very non-AI problem: while 64% of employees possessed the skills necessary to support successful change, and 74% were willing, only 25% effectively adapted their work habits. It's no surprise then that companies integrating strong change management into AI projects are 47% more likely to hit their targets. Yet, further Gartner research from 2024 revealed that a mere 25% of companies have a comprehensive change management strategy for their AI endeavours. For some reason, it seems like most organisations mistakenly believe that the biggest change in modern workplace productivity will easily be adopted with just a few words of gentle encouragement.
Another significant hurdle is the prevalent skills gap and lack of AI literacy. Many employees simply haven't been equipped with the knowledge or training to use AI tools effectively. It's telling from Wharton’s research on the topic that only 40% of executives report their organisations provide formal AI training. Insufficient training actually accounts for a whopping 38% of all AI adoption challenges in businesses, according to Mckinsey. Employees are often willing, with nearly half stating that more formal training is the single best way to boost AI adoption. Without it, they'll stick to the familiar, less daunting, old ways of doing things.
Even if an AI tool makes someone's job easier, if there's no personal benefit, recognition or reward for embracing it, some workers may not even bother. In fact, the Wharton study revealed that some workers might even fear that increased efficiency through AI will lead to a heavier workload or, worse, make their role redundant. Conversely, the right incentives can be a powerful motivator. The studies reveal that financial rewards or recognition for using AI would significantly improve uptake, but in reality, providing financial reward for individual contributors for this kind of thing can create more problems. Recognition, on the other hand, can be a compelling way of driving support across an organisation.
Finally, visible leadership support and dedicated resources, like time to experiment and access to AI mentors, are also crucial but often lacking, with over a fifth of employees reporting minimal to no organisational support in learning AI tools, particularly whilst balancing existing heavy workloads.
Ultimately, the journey to effectively harnessing artificial intelligence is far less about technological procurement and far more about navigating a complex web of human and organisational dynamics. The evidence paints a clear picture: while the strategic imperative for AI is undeniable and leadership ambition is high, a significant chasm exists between this vision and the reality of day-to-day employee adoption. This divide is not attributable to a single cause, but rather a confluence of deeply ingrained behavioural barriers, systemic shortfalls in organisational preparedness and change management, persistent skills gaps, cultural inertia, and a tangible lack of consistent, visible support for knowledge workers. Addressing these interconnected issues is therefore the critical next step for organisations seeking to move beyond aspiration and unlock the genuine productivity gains AI promises.
Strategies for successful AI adoption
It's clear that merely introducing AI tools into an organisation doesn't guarantee their use or success. While acquiring the technology can be straightforward, fostering genuine human adoption presents a more significant challenge. For organisations committed to ensuring AI initiatives are impactful and deliver tangible benefits, here are some key strategies to consider:
Establish a clear purpose, communicate broadly and frequently
Ensure your people understand the organisational and individual benefits of AI adoption through early, frequent, and transparent communication. Set company goals, and find internal AI advocates to run the hype machine and go after them.
Treat AI adoption as a change programme
Adopt a formal change management model, such as Kotter’s 8-Step Change Model, which aligns extremely well with this kind of change.
Invest in skills development & ongoing support
Provide continuous, role-specific AI training, build AI literacy, and set up mentorships for people who are struggling to adapt.
Foster inclusive participation & co-creation
Involve end-users in the AI integration process from the start, soliciting feedback and co-designing solutions to build ownership.
Support a safe-to-learn culture
Encourage experimentation and destigmatise the AI learning curve by fostering psychological safety and celebrating effort.
Provide clear incentives & recognise engagement
Reinforce AI adoption through tangible benefits, career development links, public recognition, and active managerial support.
Measure and learn
Define success metrics, track behavioural adoption, share progress transparently, and use feedback to refine your approach.
Success starts with our people
The journey to successful AI adoption is less about the tools or AI itself, and more about inspiring a transformation. The shiny new AI tool is just the starting line; the real work lies in thoughtfully, patiently, and persistently guiding your people through behavioural change. It's undoubtedly the harder part, but with a people-first approach, it’s a challenge that can absolutely be conquered, unlocking the immense potential AI can bring to our work.
