AI licences were bought, an announcement went out, maybe there was an introductory seminar. But a few months later, usage reports disappoint: a small group uses the tools heavily, most barely at all.
This is often read as employees "resisting technology". Research points to more concrete causes. This article covers the six most common barriers and a practical fix for each.
How common is it?
- BCG's AI at Work 2025 survey of more than 10,600 workers in 11 countries found that regular use among frontline employees has stalled at 51%.
- In the same survey, only about a third of employees consider their training adequate.
- According to Microsoft and LinkedIn's 2024 Work Trend Index, 75% of knowledge workers use AI at work, but 78% of AI users bring their own tools. In other words, use often happens where the organisation can't see it.
So the problem isn't just "not using AI"; it's not using approved tools under the organisation's rules.
1. Unclear rules: "Am I allowed to enter this?"
When employees don't know what data may go into which tool, two things happen: the cautious don't use AI at all, the bold use it riskily. Neither is what you want.
Fix: A short, clear AI acceptable use policy. A good policy is less a list of bans than a clear guide saying "these kinds of work are fine with this tool". Türkiye's data protection authority's 5 March 2026 announcement on generative AI at work also stresses corporate policies and informing employees.
2. Training that's too little or too generic
A one-off "what is AI" presentation doesn't tell employees how to apply it to the work on their desk on Monday morning.
Fix: Training specific to roles and units, built on their own work examples. In BCG's research, 79% of those who received more than five hours of training were regular users, compared with 67% of those who received less; in-person training and coaching strengthen the effect.
3. Managers don't guide
If a manager doesn't use the tool, doesn't tell the team to try it and doesn't allow time to experiment, employees read it as "not a priority". In BCG's survey, only 25% of frontline employees say they get enough guidance on AI from their managers.
Fix: Separate training for managers, and a regular "what did we try with AI this week?" slot in team meetings. A manager describing their own use is a strong signal.
4. No bridge between the tool and the work
An employee opens the tool, faces an empty chat box and doesn't know what to type. After a few attempts that don't give the expected result, they give up.
Fix: Ready-made use cases and an example prompt library for each unit, with concrete starting points such as "Summarise this customer complaint in this format" or "Interpret this table in management report language". Prompt engineering techniques help here.
5. Worries: will I lose my job, will I look lazy?
In Microsoft's 2024 research, more than half of AI users said they don't want to admit using it at work for fear their employers might see them as "lazy" or "replaceable".
Fix: Open communication. Say clearly why AI is being used (to lighten which work), where human judgement stays in charge, and that use is valued. Making good examples visible, for example through monthly "best use" sharing, reduces hesitation.
6. No support after training
Enthusiasm is high the week after training; a few weeks later, old habits return. Employees who get stuck have no one to ask.
Fix: AI champions in each unit and regular short Q&A sessions. Champions are volunteer early adopters who support colleagues in their daily work. We explain how to set them up in our article on building an AI champions network.
Which barrier dominates in your organisation?
| Symptom | Likely barrier |
|---|---|
| Frequent "am I allowed to enter this?" questions | Unclear rules |
| A short spike after training, then a decline | Lack of support |
| Very high use in some teams, zero in others | Manager effect |
| Widespread use through personal accounts | No approved tool, or an approved tool that isn't useful |
| Use exists but is always the same simple tasks | Generic training, missing use cases |
Conclusion
Employees not using AI is usually not an attitude problem but a design problem: it appears when rules, training, manager support and continuity aren't addressed together.
To see your organisation's usage risk, take the free GenAI Usage Risk Assessment. For a process that tackles all six barriers together, starting with an assessment and ending with a results report, explore our Corporate AI Adoption Programme.
Sources
- UNLEASH – Four takeaways from BCG's AI at Work 2025 report
- Management-Issues – BCG: Move beyond deployment to unlock AI's full potential
- Microsoft – 2024 Work Trend Index announcement
- KVKK – Use of Generative AI Tools in Workplaces (5 March 2026, Turkish)
- Prosci – Applying ADKAR to AI adoption challenges