What our AI bootcamp taught me, and what I would do instead
In mid-2026 I sent my whole team through an AI bootcamp. It cost money and changed very little. Here is what went wrong, and the simple four-step way I would start today.
I want to begin this journal with a mistake, because I believe mistakes teach more than success stories do.
Like many people, I started trying AI chat tools in 2025. In early 2026 I tried to bring AI into our business, but I did not have a clear plan. I knew it mattered. I did not know where to start.
So in the middle of the year, I did what felt right at the time. I organised an AI bootcamp for my whole team. Everyone would learn the tools together, and then, I hoped, everyone would use them.
It did not work out that way.
What happened
The bootcamp itself went well. My team was curious, asked good questions and tried things. But when everyone went back to their desks, very little changed. The work came back, the deadlines came back, and the new tools sat unused.
We spent money on training and on tools that we did not really need. Looking back, the cost was not only the money. It was also the time of a whole team.
It is easy to leave a session like that thinking that AI is interesting, but not for your own job.
What I got wrong
I put the tools first and the work second. We trained everyone on many tools before we had chosen one real piece of work for those tools to do.
A tool with no job is just another app on your phone. People learn by doing real work, for a real reason, not by watching a demonstration.
I also made it everyone’s job, which meant it was nobody’s job. No single person was responsible for making one thing work.
What I would do instead
If I could start again, I would not begin with a bootcamp. I would begin with one task. This is the simple method we follow now:
- Choose one weekly task. Something real that happens every week and takes time: a weekly report, follow-up messages after a quotation, replies to common customer questions, or a plan for your social media posts.
- Give it one owner. One person tries AI on that task and is allowed to make mistakes. She is the one who learns what works.
- Measure the time saved. Before you start, write down how long the task takes. After a few weeks, write it down again. Keep it simple: minutes on a piece of paper are enough.
- Then widen. When it works, the owner shows the next person, using a real example from her own work. Then you choose the next task.
One more thing: choose the tool after the task, not before. Most of the time, the tool you already have is enough to begin.
Why this works better
One task is small enough to finish. A real task gives people a reason to keep going on a busy Monday. One owner means someone cares whether it works. And measuring the time turns a feeling into something you can see.
The best part is how it spreads. When a colleague sees her teammate finish the weekly report in far less time, she does not need a bootcamp. She simply asks her teammate to show her how. That is how real change moves through a team: from one person to the next, with a real example.
A few safe habits from the start
Whatever task you choose, start with good habits:
- Do not paste private details about clients or staff into a tool you have not checked.
- Read and check everything the AI writes before it goes out. It drafts; you decide.
- Keep your own voice. Change the words until they sound like you.
If you lead a team
If you lead a team, my advice is simple. Keep the money you would spend on a big training day. Instead, spend a little time choosing one task, one person and one simple measure. Training can come later, when people have a real reason to want it.
Our team did find its way in the end. I share how we moved from chat tools to building our own tools in another journal entry.
Mistakes are part of building. I am sharing this one so that you can skip it.
Not to replace her. To empower her.
