The problem
Employees use AI tools at work, even when the organisation has not approved them: to write texts, analyse data or summarise emails. This is called shadow AI. In a UK survey commissioned by Microsoft (2025), 71% of employees said they had used unapproved AI, and about half did so weekly. More than a fifth reported using a consumer AI assistant for financial tasks.
Why it keeps coming back
- The tools are free and immediately available.
- Employees simply want to get their work done faster.
- The organisation has no equally good alternative of its own.
The risks
- Company and personal data end up with outside parties.
- Nobody knows which outcomes were produced by AI and whether they are correct.
- Knowledge and ways of working sit in separate tools instead of in the organisation.
Our approach
- Do not ban, offer something better. An internal AI service that is at least as easy to use, connected to your own documents and systems, so it can do more than a stand-alone chatbot.
- On your own hardware where needed. Sensitive data stays inside.
- House rules. Short and clear: which tool for what, which data is and is not allowed.
- Visibility. What the AI does and proposes is recorded, so outcomes can be traced.
What to watch out for
- Ask employees what they use AI for now. That shows you the need straight away.
- Make the safe route the easiest one.
- Invest in AI literacy: people who know how AI works make better choices.
Would you like advice on AI use by employees? Request a free advice conversation.
Sources
Updated on