How do you know your team is ahead of you?
You rarely hear it said. You notice it. A draft arrives faster than the person could have written it and you cannot tell what is theirs. Someone uses a term you nod along to. A report lands in a shape you do not recognize. You are asked a question about a tool and have no view. None of this proves anything, but together it is a prompt to check.
The wider evidence points two ways, and it is worth holding both. On one side, hidden use is common. Microsoft and LinkedIn's 2024 Work Trend Index, a survey of 31,000 people in 31 countries, reported that 78 percent of AI users bring their own tools to work. The University of Melbourne and KPMG 2025 global study, of more than 48,000 people, reported that 57 percent of employees admitted using AI in non-transparent ways, such as presenting AI-generated content as their own, and 48 percent said they had uploaded company information into public AI tools.
On the other side, the average leader is not behind. Gallup's fourth quarter 2025 U.S. survey, published January 25, 2026, found that 44 percent of leaders use AI at work a few times a week or more, against 30 percent of managers and 23 percent of individual contributors. So "my team is ahead of me" can be true for a specific person or team in your company and false as a general rule. Do not assume either.
What should you not do?
- Do not make your first experiments a public performance. In my view a leader who asks the team to teach them in an open meeting, or announces an AI plan before they can judge one, hands the room a version of them with no judgment yet. Learning in front of staff has a place later, once you have something to show.
- Do not use shadow tools. If you paste company material into a personal account, you do what you would not want staff to do. KPMG's study found that uploading sensitive information was most common where organizations had banned generative AI (67 percent) and still frequent where a policy existed (56 percent), and its authors suggested outright bans may be ineffective. In my view clear guidance works better, and it starts with you using approved tools.
- Do not write policy first. A policy drafted by someone who has not used the tools tends to ban what is harmless and miss what is risky.
- Do not interrogate the person who is ahead. Ask them how they use it, out of curiosity. People who feel audited stop talking.
What is a first-week plan you can do alone?
Seven days, about thirty minutes each, on your own. Use an AI assistant your company approves, or your own paid account for material that is not sensitive. ChatGPT, Gemini and Claude are the usual starting points.
- Day 1. Pick one low-sensitivity task from your real week, such as summarizing a public report you already know well. Run it. Do not judge the tool yet.
- Day 2. Run the same summary again with a clearer instruction, including who the reader is. Compare the two answers.
- Day 3. Ask it to rewrite one of your own emails. Notice what it gets wrong about your voice.
- Day 4. Ask it to prepare you for a meeting using only information you would be comfortable sharing publicly.
- Day 5. Ask it a question you already know the answer to in your field. Note every place it was wrong or too sure. Keep this list.
- Day 6. Ask the person on your team who is furthest ahead one specific question about how they use it.
- Day 7. Decide one routine worth keeping and one rule for confidential material, in writing, for yourself.
The aim is judgment, not skill. The U.S. NIST generative AI profile names confident but false output and over-reliance as risks, and your Day 5 list is the best protection you can build.
What does catching up not require?
It does not require code, a course or the same fluency as the most advanced person on your team. It requires enough use to ask good questions, spot a weak answer and say which material may go where. That is a smaller target than it feels like.
When does private coaching help, and when does it not?
It helps when you want to work on your own documents and tasks, with no audience, and want a second pair of eyes on the output. I coach executives one to one on video, building tools on their real material, as set out on the coaching page and in the session walkthrough.
It does not help if you have not tried the tools yet, so do the week first. It does not help if the real gap is company-wide ownership of AI, which is an operating role, covered in the decision guide and on the CAIO page. And it will not work if you want it done for you, because the point is that you can judge the output yourself.
Sources
- Microsoft and LinkedIn, 2024 Work Trend Index (May 8, 2024). Survey of 31,000 people in 31 countries. 78 percent of AI users are bringing their own AI tools to work.
- University of Melbourne and KPMG, Trust, attitudes and use of artificial intelligence: A global study 2025 (May 2025). More than 48,000 people in 47 countries. 57 percent of employees admitted non-transparent AI use, 48 percent reported uploading company information into public AI tools, and the behavior was most common where generative AI was banned (67 percent).
- Gallup, Frequent use of AI in the workplace continued to rise (published January 25, 2026; survey October 30 to November 14, 2025). U.S. employees. Frequent use (a few times a week or more): leaders 44 percent, managers 30 percent, individual contributors 23 percent.