Hello Reader !

​

End of June! Besides, this is my birthday month 😉, and it also marks the end of the school year. And time to reflect on what my kids and I learned this year. We all had to learn GenAI. So this made me realize that if I were to open an "AI school for leaders", I would focus on helping my students to embrace only one shift.

AI is not a tool; it is a team.

Most leaders think GenAI adoption requires learning entirely new skills. Wrong! If you can hire, onboard, and manage human teams, you already have 90% of what you need.

The problem? Organizations view AI as a technology transformation rather than replicating an existing operating model they have already mastered. This mindset shift is at the core of what enables some leaders to see AI's potential, while others remain stuck.

What you'll learn today:

• Why your current leadership skills translate perfectly to AI
• A 7-step framework to build your AI team (just like hiring humans)
• How to avoid the complexity trap that stops most teams

​

​

Get comfortable for your 5 minutes of sparks to hack.

​

✨ Spark

​

✏ Stop Treating AI Like Technology

Most people see GenAI as a technology challenge or digital transformation project - like adopting the internet was 20 years ago. They think they need to become AI experts first.

But AI is not a TOOL, is it a TEAM.

​

✏ AI is not a tool, it is a team

No, treating it as a human team. But treating like a human team, so applying your best leadership style. It's about building and managing a team - something you already know how to do. You don't need technical skills. You need the same leadership skills you use every day: defining roles, setting expectations, giving feedback, and optimizing performance.

I'll show how.

​

✨ Hack

Let's break down "AI adoption" into the 7 steps of hiring a new AI team member:

✏ Step 1: Define your job profile

Start like any hiring process. What's your vision and objectives? What blocks you - time, skills, motivation? Map your current strengths and identify gaps. This is where your awareness and honesty will make a significant impact.

Then define what type of AI expertise you need. For exemple recently, I hired NEA, a negotiation and assertiveness expert, to help me build this competency.

​

✏ Step 2: Start the hiring process

​

Pick one LLM (ChatGPT, Claude, Mistral, or Gemini .. ) and start interviewing. Test different prompts like you'd test different candidates. See how the AI responds to your management style, your detailed instructions, and open questions.

​

✏ Step 3: Collect the onboarding documentation

​

Just like onboarding new hires, your LLM needs SOP, templates, and frameworks.

Standardize your data and collect all the knowledge an LLM needs to perform the task in a central location.

HUB, help me to translate a newsletter into LinkedIn posts(I always write the source document, I love writing, and don't want to lose that skill 😅). Therefore, I needed to provide HUB with numerous templates of high-quality LinkedIn posts.

For you, these could include templates for a strategy summary, internal or external memo, or a pricing negotiation position paper. Also, customize your LLM to your writing style and tone should be part of your onboarding process.

​

✏ Step 4: Run a trial period

​
Run a probation period - but faster. Instead of 3 months, give it 1-2 weeks.

Test interactions with a detailed prompt (I like to have a shortcut of my prompt library to limit friction). Then monitor effectiveness. Does this LLM + your prompt + the information you gave deliver what you need? If not, adjust or "fire" it.

​

✏ Step 5: Create a dedicated Assistant

When you're confident about the quality of the work and the candidate, formalize the role in your AI team. Create a dedicated assistant (GPT, Claude Project, Gem in Gemini or Copilot agent). Give it a clear job description and consistent responsibilities.

I wrote a newsletter about how to create an assistant, step by step, for a non-techie like me here.

​

✏ Step 6: Give Feedback

​

Provide regular feedback like any good manager. When outputs aren't right, explain what you want differently. You can also share examples of good work.

Constantly update instructions based on what works; it is called reverse engineering. It means once you are happy with the output from your assistant, you ask your LLM to send back the updated instruction. So next time, you will get directly to this final output, saving you a lot of time and energy 😡😜.

​

✏ Step 7: Develop and train your AI team

​

Don't settle for initial performance! You need to keep developing your AI team.

The performance of the LLMs evolved rapidly so maybe something that was not possible before is now easy: for exemple you can link ChatGPT to your G drive or Claude has now Internet access. These updates might impact the LLM selection or the potential of delegation of some tasks. It is like having a team member sent to training without you been informed - check their competencie regularly.

Also be disciplined in updating their knowledge (aka the data and context you provide) constantly.

As your AI team grows, (I have almost 30 assistants), make sure to adapt roles and instruction to avoid duplication (yes this also exist in the AI world 😅). Also retire assistants you don't use anymore. I planned to do that this summer to simplify my AI team structure.

​

That's it for today.

Reader, hit reply to let me know if you liked this edition or if you tried these tips. I respond to every person who writes to me!

See you in two weeks. Keep the spark alive, and be intentional :)

What's up with me?

I have updated my website! The first one was done by me (2019), the second version by a friend (2022), and now I have hired a professional webmaster and designer. Still need to be improved, but this is much nicer. Let me know what you think. We also updated the new dates of our ScaleU with AI coaching program. It is here 👉 www.bambooster.ch.

​

​