A PwC study shows that 95% of Danish business leaders use or plan to use AI within 18 months. At the same time, half admit their company has a low competence level in the area.
This is a dangerous combination: high ambition and low readiness. And it explains why so many AI projects end as expensive pilots that never become operational.
The problem is rarely the technology. ChatGPT, Copilot, and Claude are accessible, powerful, and relatively inexpensive. The problem is that companies treat AI implementation as a technology challenge when it is actually a change management challenge.
Five barriers that slow AI adoption
Based on experiences from Danish companies and international research, we see five recurring barriers:
- • Lack of trust: Employees doubt AI output, especially when they do not understand how the model reaches its conclusions. Skepticism is healthy — but without addressing it, it becomes blocking.
- • Underestimation of potential: Many only use AI for simple tasks (summarization, rewriting) and see it as a 'fancy search engine'. They never discover the transformative use cases.
- • Fear of job loss: Implicit or explicit fear that AI will make employees redundant. This fear is often rational — and must be taken seriously, not dismissed.
- • Lack of competence: AI tools require new skills — prompting, critical evaluation of output, understanding limitations. Without training, tools remain unused.
- • Time pressure: Paradoxically, employees 'don't have time' to learn tools that would save them time. PwC estimates that up to 40% of working time is spent inefficiently on routine tasks.
The CEO insight: AI saves time — but not the bottom line
A Danish CEO put it precisely: many companies do not get financial benefit from AI's time savings because employees simply work less covertly.
It is an uncomfortable truth. If AI frees up 10 hours per week for a department but those 10 hours are not allocated to new tasks, the company has invested in AI without return. The technology works — but the organization has not adapted.
The counterexample is companies that deliberately design AI into their operations. Dinero, the Danish accounting software company, froze hiring and required everyone to find AI optimization solutions. The result: the same number of employees serve far more customers with the same customer satisfaction. It requires leadership will — not just technological capability.
Why change management is more important than technology choice
Over 90% of companies that have succeeded with AI implementations spent half their budget on implementation — not on technology. That number should stop any leader in their tracks.
AI implementation is a change process. And change processes succeed when three things are in place:
- • Leadership anchoring: Leadership uses AI visibly and communicates clearly why it is a priority. Not as hype, but as a concrete tool for achieving business objectives.
- • Employee involvement from the start: Those who will use AI should help define what it is used for. Super users and champions serve as catalysts — but only if they have leadership backing and time.
- • Tailored use cases: Generic AI training ('Learn to use ChatGPT') has limited effect. What works is use cases close to employees' daily work: 'How to use AI to prepare the weekly sales meeting.'
Our approach: 6 steps from pilot to anchored operations
At Vertex Solutions, we have developed a 6-step method that systematically addresses the human dimension of AI implementation. We do not treat AI as an IT project — we treat it as an organizational change.
The method ensures we do not just build the right technology, but that the technology is anchored with the people who will use it. This includes needs assessment with employees, pilot testing with real use cases, governance frameworks, and ongoing evaluation of adoption and satisfaction.
We have seen the same pattern repeatedly: companies that invest in change management parallel to technology development achieve 3-5 times higher adoption rates than those focusing only on the technical.
It is not about choosing between technology and people. It is about understanding that technology only works when people are on board. Read more about our ethical approach to AI implementation that puts employees at the center from day one.
Conclusion: Invest in people — not just in models
AI models get better every month. Technology is not the bottleneck. The bottleneck is the organization's ability to absorb change — and that requires deliberate, strategic work with culture, competencies, and leadership.
If your AI pilot has stalled, or your employees only use ChatGPT to rewrite emails, the solution is not a better model. It is a better implementation process.
- • Spend half the AI budget on implementation — not just technology
- • Address the five barriers directly: trust, understanding, fear, competence, time
- • Design AI into operations — do not let time savings disappear covertly
- • Create use cases close to employees' daily work — not generic training
- • Involve employees from needs assessment to operations
- • Measure adoption and satisfaction — not just technical performance