Hello Reader !

I am preparing for my International Coaching Federation exam (ICF) next week. As a system coach, you learned and get assessed on the power of your questions. So, how do we use questioning with GenAI? Most people use it like a search engine (Ecosia, Google). They ask questions and expect instant answers. But the real power comes when you flip the script and let the GenAI question you first.

What you'll learn today:

Why does choosing the direction matter more than speed? So why do you need to get asked questions by your LLM?
How to use GenAI with a questioning step.
Technical tips for the lazy (like me 😉) who don't want to spend too much time answering questions or writing long prompts

Before we begin, how can I help you?

Help you build an AI system mindset: If you are about to shift career, leave your corporate job, or set up a coaching and consulting company, this is what you need here.
Coach you and your team, so we can design and pressure-test your new strategy (e.g., vaccine for cancer) or product (e.g., AI& digital products) faster.
Audit and improve your innovation, customer centricity, and ability to adapt to a complex and ever-evolving environment: I will train your teams on GenAI, Customer-centricity in healthcare, intrapreneurship, and agile & startup ways of working.

"Judge a Man by his questions rather than his answers." - Voltaire"

You can listen or watch the content of this newsletter here 👇.

Get comfortable for your 5 minutes of sparks to hack.

✨ Spark

✏ GenAI is not Google search

People think GenAI's value is in fast answers. But speed without direction is just busy work.

We're falling into the "fast and pleasing mode" of LLMs – getting one answer quickly instead of exploring multiple options and perspectives.

Common mistakes:

Using AI like a search engine instead of a thinking partner
Jumping straight to asking for solutions
Accepting the first GenAI response without exploration

✏ Let your LLM ask you more questions

If, you want to improve your output, start there. Ask your LLM to ask you questions:

It helps you pause, so that often reflective questions emerge: Do I really need to do that? Should I delegate, or stop?
It helps you to clarify your goals and the scope.
It helps you expand your thinking: you will explore the problem or the context you want to get an answer for.

✏ Fight against the fast and pleasing mode of your LLM

Forcing your LLM to ask you questions before it answers will help you to fight the fast and pleasing attitude of LLMs.

Fast - getting 2 pages of answers in 5 seconds.
Pleasing: as it learned what you expect (eg if you use the memory option in ChatGPT, your LLM remembers previous chat and your feedback). You will rarely get a contrarian view or surprising feedback.

It's about cognitive expansion (cognitive decline is also a risk with GenAI 😅😅 - but not the scope of today's discussion). We tend to think fast and narrowly when GenAI can help us think deeper and broader.

✏ Direction matters more than speed

This quote from Naval’s book, The Almanack of Naval Ravikant, is one I’m using each week in coaching (by the way, I highly recommend the book ❤️).

Picking the direction that you’re heading in every decision is far, far more important than what force you apply. Naval Ravikant.

Back to GenAI and the value of questioning. It helps to decide the direction instead of being guided by your GenAI. It also prevents the misuse of resources and energy on useless output.

Now I hope you are convinced why we should do it - but this questioning could slow down the process. But that's the point!

I will show you how to make it very easy and fast for you.

✨ Hack

✏ Get a template prompt with step 1: ask me questions

We are all lazy, so to avoid skipping this step, add that in your prompt template as a mandatory step.

Your prompt template is the one you copy and paste each time. You can copy mine: In the Process Instruction section of my prompts, I have "always start by asking me at least 3 clarification questions so that you can do your work the best".

Create your library of prompts: join our ScaleU with AI coaching program

✏ Use a microphone if you're too lazy (like me) to type

The voice mode (aka the microphone to dictate to your LLM) is definitely the feature I've been using the most, and I see it with my clients too.

Just last week, I was coaching on Copilot, and using the mic helps people add more context and get more questions answered. It ultimately improves the output (with little effort).

A client of mine was looking for an 80-year-old gift for his grandma. Instead of "What gifts for a 90-year-old woman?" he asked "Ask me questions to understand what would make a meaningful gift for my 80-year-old grandma." The GenAI asked about hobbies, recent life changes, relationship dynamics, and her personality type.

The value of a question is not in the answer, but in the thinking it provokes.

✏ Choose your angle of the questions

Choose an expert: the role you want GenAI to act as

Tell the GenAI to question you as a specific expert would. "Question me like a productivity optimization expert", "Interview me like a senior team coach", "Interview me like a GenAI consultant", creates deeper, more relevant inquiry.

Recently, I was designing a product vision template for a diabetes application. I asked my LLM to act as a Senior digital product management expert in life sciences to scope what needed to be covered.

Get interviewed by a famous person:

These tips make you feel famous 🤩, as if a famous person interviewed you. For example, if you love Simon Sinek and have to come up with a solution for your team, you can have it act like Simon Sinek interviewing you. Or, for me, I love the content of Brice Challamel on GenAI (head digital at Moderna), I often use that to clarify a situation with GenAI.

I hope you find these tips inspiring. Reader, let me know if you use some.

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

What's up with me?

Last week, I led the opening session of "Discover your cohort" for the DayoneHealthtech acceleration program here in Basel. We had 15 TechBio and HealthTech ventures developing AI solutions to enable Pharma R&D. Such a smart group. I made them sketch and present their solution ( Who benefits, How it works, What is the impact), and the main risks and opportunities. This was fun 😉 and created trust and engagement for the rest of the day.

Spark to Hack ✧