Why are so many organisations struggling with AI adoption? Why aren’t the much-lauded productivity and efficiency gains materialising? The answer isn’t technological. It’s cultural.
While substantial investment has been made in AI tools, far too few businesses have made a parallel investment in upskilling their people, to successfully embed AI into workflows.
Is workplace AI adoption about to plateau?
In May, technology journalist Azeem Azhar drily observed that AI backlash is the only thing currently growing faster than AI revenues [1]. He catalogued a range of instances where the Silicon Valley behemoths have been publicly criticised for prioritising profits over people, and more broadly, humanity. Yet the issue is more deeply nuanced than attention-grabbing media headlines show.
One of Gen Z’s many stereotypes is that, as a generation, it’s hyper-connected and pro-tech. But rather than blithe adoption, young people are showing themselves to be acutely aware of AI’s shortcomings. Hardly surprising when the number of graduate jobs is in sharp decline [2] and AI is recognised as a contributing factor. Or when AI data centres are shown to carry a colossal environmental footprint [3].
Investment in AI has far outstripped comparable investment in associated upskilling and reskilling so it’s little surprise that AI adoption is proving problematic. It’s one thing to purchase new technology but it’s something else entirely to make sure an entire workforce is competent and confident in how best to use it.
And even when such investment exists, the data shows AI productivity gains are slow.
In April, Accenture reported [4] that despite a sharp rise in AI use, only one in ten UK firms has yet successfully scaled it or embedded it into core workflows. Its Generating Impact report anticipates that long-term economic value will derive from AI-fuelled higher-quality outcomes, not cost efficiencies. Yet this value won’t materialise when people feel threatened and uncertain.
Successful integration of enterprise AI is starting to look far more like a culture issue than a technology one. As such, it’s worth dissecting what needs to happen for people at work to start unlocking this technology’s full potential.
We’ve identified five things people need, to make progress with new technology. In the aggregate, they present a compelling case for prioritising workplace culture if organisations hope to successfully integrate AI.
Let’s look at each in turn.
1. Are your people psychologically safe?
Without psychological safety, fear and anxiety can easily spiral. This shuts down the cognitive reasoning that thoughtful AI adoption requires.
Fear is a primal human emotion and natural response to perceived danger. But threat changes how the brain operates. Fear activates the amygdala (the emotional response centre and what Dr. Steve Peters calls the ‘chimp brain’ [5]), shutting down the prefrontal cortex (responsible for focus, emotional regulation, reason, problem-solving and critical thinking). It triggers the fight, flight or freeze response.
While this is a powerful defence mechanism in the face of a genuine threat to life, in other situations it is debilitating. Perceived threat reduces the ability to respond proportionately.
AI-related job losses are widely documented in the media and in many instances is exaggerated. Bad news stories spread far faster than good news and media outlets are under ever-increasing pressure to hit circulation targets. In the absence of alternative narratives, it’s completely understandable that people feel anxious about the possibility of their livelihoods being undermined by AI and automation.
How can businesses address AI job loss fears?
Any employer investing in AI must be willing to create a vision of the future that positions these technologies as an enabler rather than a replacement. No one can be expected to learn how to use AI optimally if they fear the long-term implications.
Alongside, the ethical issues arising from AI use can’t be underestimated. As both economic and environmental conditions worsen, it’s entirely reasonable that many will hold strong views on the wider consequences of unregulated AI adoption.
Taking time to understand and address peoples’ concerns regarding AI is a vital first step towards lasting adoption.
2. Are you telling the right story about AI?
Compelling narratives make abstract AI concepts accessible and emotionally resonant. This underpins lasting behaviour change.
Humans find and create meaning through story. They are far more memorable than data points and storytelling is a powerful tool for helping people change.
Research shows emotion drives lasting behaviour change far more effectively than logic. When people hear a story of how someone achieved something (how they integrated AI into a workflow, for example), the idea of doing something similar immediately becomes far more accessible.
Stories help us create new identities, showing change as possible. Hearing how someone else overcame obstacles to make progress is motivating and affirming.
Stories also build trust. Sharing a story creates a sense of ‘we’ and ‘us’, fostering a sense of community. This increases the likelihood of lasting behaviour change as people are far more likely to shift habits when they feel connected to a greater whole.
What makes an effective AI narrative?
A compelling AI narrative that sets out a clear path for adoption and minimises the risk of unforeseen consequences. It instils a sense of safety and security. Just as a well-crafted strategic narrative fosters shared understanding, fuels engagement and creates organisational cohesion, a clearly articulated vision of the role AI will play in people’s work futures builds courage and confidence.
3. Are you allowing people time to play?
Unstructured learning time and low-pressure experimentation build confidence and competence.
Today’s work environments are ambiguous and harried. Continuous change is the new normal and yet rising geopolitical and economic uncertainty make it increasingly hard to make plans with any degree of certainty.
Against this backdrop, AI has been positioned as a panacea to improve efficiency and ergo, bottom-line outcomes. But these technologies are evolving at lightning speed and in the rush to make gains, too few organisations are prioritising time to reflect or make sense of where things are at.
Learning how to use AI safely and responsibly requires ringfenced learning time. It requires time to get familiarised with new tools, processes and practices. Setting time aside to play has huge benefits.
The British Psychological Society sets out three key attributes to play [6]. It must be enjoyable; voluntary; and done for its own sake. This is vital, yet sadly most of us don’t associate play with work.
Play allows experimentation without pressure, which reduces fear of failure. When there’s no pressure to get things right first time, people are naturally more likely to try things they otherwise wouldn't. Experimenting without pressure is enjoyable, increasing the likelihood of exploration and trying out new things. It leads to more time spent honing skills that build mastery.
From a learning perspective, unstructured play and exploration helps the brain to recognise patterns faster as lived experience always surpasses theoretical instruction. In a safe and unhurried environment, mistakes become feedback rather than failure. This one shift alone deeply enhances both engagement and competence.
In short, play boosts both confidence and competence and embeds learning.
4. Are you setting clear boundaries?
Clear governance frameworks provide the guardrails and reassurance people need to engage with AI confidently.
As AI evolves, clear governance frameworks are needed for its safe and ethical integration into workflows. New processes, rules and practices are a standard feature of successful business transformation and AI adoption is no different.
As AI providers add new features on an almost weekly basis, governance frameworks should be regularly reviewed and updated by a multidisciplinary taskforce that thinks holistically about how to reduce risk and ensure safe enterprise-wide AI adoption.
What should an AI governance framework include?
Asides from providing essential organisational structure, colleagues need frameworks that provide clear rules for engagement. They need to know what they can do with AI and what they can’t.
Without this, given such economic uncertainty, people will opt for doing nothing rather than run the risk of doing something that later threatens their job security.
An AI governance framework provides the much-needed oversight that prevents misuse and reduces the risk of unintended consequences.
5. Are your people empowered to say no?
Critical thinking and the confidence to challenge AI outputs are essential safeguards against errors and misinformation.
More than anything, people need the confidence and courage to challenge AI outputs. This requires discernment.
Despite progress being made, large language models (LLMs) are still regularly prone to hallucinating [7]. When data and information sources are sparse, AI systems are far more likely to generate responses that sound plausible but are in fact inaccurate. Because GenAI outputs regularly sound authoritative, without critical thinking, it’s easy to assume such sophisticated technology has access to superior knowledge.
But it’s vital to recognise that online misinformation and disinformation are rife. Indeed, the World Economic Forum ranked disinformation as the fifth greatest risk globally in its 2026 Global Risks Report [8].
We mustn’t overlook the reality that the quality of LLM output depends entirely on the quality of the datasets it’s trained on. As the computer science saying goes, “Garbage in, garbage out”.
Being vigilant to the current limitations of these emerging technologies is key. But linking back to the previous four points, vigilance requires reasoning. Which first requires psychological safety.
The First Step: Conduct a Culture Audit
Before investing further in AI tools, organisations should be able to answer the questions: Do our people feel psychologically safe? What narratives are already circulating? Do people have time to experiment? Do they understand the rules of engagement?
Making this assessment will reveal where adoption is actually breaking down—and why.
Conclusion
The case is clear. Successful AI adoption isn't a technology problem. It's a people problem.
When colleagues lack psychological safety, lack coherent narratives, lack time to experiment, lack clear boundaries or lack the confidence to challenge LLM outputs, AI adoption will stall.
The organisations that will unlock AI's full potential aren't necessarily those with the biggest budgets or the boldest ambitions.
They're the ones most invested in nurturing healthy and inclusive work cultures. When people feel informed, empowered and supported, they will naturally feel more inclined to lean into innovation and change.
The good news is that any work culture can be adapted. But this requires honesty, commitment and deliberate, persistent action.
How Working the Future can help
We’re helping organisations understand exactly how and where their existing workplace culture is impacting AI adoption. And then we’re helping create the conditions for adoption to thrive.
Our culture audits identify what's going on beneath the surface: the fears, the narratives and the gaps in confidence that quietly hold people back.
Our tailored workshops give colleagues the psychological safety, courage, confidence, clarity and space to engage meaningfully with AI.
Our discovery sessions bring teams together to build a shared vision and governance framework that the whole organisation can get behind.
Internal communication support ensures the entire process is communicated well, to create a coherent, compelling narrative that makes change stick.
If you want AI adoption that's sustainable, not superficial, the place to start is your culture.
Get in touch with Working the Future today.
References:
[1] https://www.exponentialview.co/p/the-ai-backlash-is-the-only-thing
[3] https://www.climateimpact.com/news-insights/insights/carbon-footprint-of-ai/
[5] Peters, S. (2012) The Chimp Paradox: The Mind Management Programme for Confidence, Success and Happiness. London: Vermilion
[6] https://www.bps.org.uk/psychologist/golden-age-play-adults
[8] https://www.weforum.org/publications/global-risks-report-2026/






