November 27, 2025
AI Is Everywhere Except in Staff Training. It’s Time to Catch Up
You can’t walk into a workplace today without spotting AI somewhere in the background. Someone’s prompting a model for a quick draft, while others are summarising a spreadsheet. Customer service teams are relying on AI assistants to speed up replies.
AI has become part of the day-to-day flow of work, but the one thing missing is the competence to use it properly. The tools are everywhere. The training has yet to follow.
The Adoption Surge and the Skills Vacuum Behind It
The majority of full-time employees use AI tools during the workday, but only 33% have received any training at all on how to use them safely or correctly.
The world’s largest AI workforce study, which includes 48,340 workers across 47 countries, reveals the same story but on a bigger scale. According to KPMG and the University of Melbourne:
- 58% intentionally use AI at work,
- only 47% received training,
- 66% don’t check outputs for accuracy,
- 48% have pasted company data into public AI tools, and
- 57% hide their AI use from employers.
In short, people are using AI to do meaningful tasks without the skills, rules, or confidence to do so responsibly.
The Risks Are Piling Up
The absence of structured training introduces new vulnerabilities into daily operations. Unverified answers become customer-facing, and sensitive data ends up in public tools. Employees over-rely on AI in areas where judgement matters. When something inevitably goes wrong, no one admits AI was involved.
Research from Calabrio shows that frontline agents are increasingly relying on AI tools while simultaneously feeling underprepared and unsure about their long-term progression. The consequences are rising stress levels and limited development support, among other things.
Without proper training, AI becomes a double-edged sword. Improper use of AI can cause inaccurate responses to slip into real customer interactions and sensitive information to be mishandled.
On the flip side, companies that invest in structured AI training are seeing a measurable difference. A recent example comes from GoHealth, a major US health-insurance provider, which implemented AI role-play software to train its sales teams. Through simulated conversations, employees were able to practise pitches, receive instant feedback and build confidence before speaking to real customers.
The approach shows what’s possible when AI is paired with proper learning design, such as structured practice, clear objectives and significant performance improvements.
Why Exactly Are Companies Not Training Their Employees?
Part of the problem is speed. AI tools are rolling out faster than businesses can design the training to support them. A 2025 Adecco survey reports that 60% of leaders expect their employees to update their skills for AI, meaning many organisations assume staff will handle reskilling on their own.
Another issue is the illusion of simplicity, because the tools feel intuitive, so organisations assume employees can simply “figure it out.” Still, ease of use doesn’t equal good judgement.
Finally, there’s the surge of “shadow AI”: employees secretly using public tools because official ones aren’t provided. The KPMG study shows how widespread this has become, with workers copying data into whatever model is easiest to reach.
What Effective AI Training Actually Looks Like
The companies getting this right are rethinking AI training entirely. Like GoHealth, they’re leaning towards role-based learning:
- Contact centre agents get training rooted in real customer scenarios
- Marketers learn prompt quality control and brand alignment
- Operations teams learn how to pair AI outputs with human judgement.
Alongside that, they’re building governance that’s clear and easy to follow, not a PDF policy no one reads. They’re offering supervised hands-on coaching, encouraging experimentation, and treating AI skills like any other essential workplace competency.
The debate isn’t about adopting AI anymore. That ship has already sailed. What matters now is whether organisations give people the skills to use it well or leave them to improvise.
