Hello Reader !

You know the scene. You are the expert in the room (eg in access, finance, legal..). A colleague arrives with an answer from ChatGPT. It looks structured, confident, polished. And it is wrong. Or half right.

It reminds me of doctors when patients started to arrive with their internet research. Consultations got longer and more complicated.

So for today:

• How AI can expand your expertise
• What's good and bad about that
• Four tips you can implement today: two for you, two with your team

Before we begin, how can I help you?

• Coach you to build your team GenAI system: skills, mindset and AI/human organization here​​
• Coach you and your team to design and pressure test your new strategy or product here​
• Audit and improve your innovation, customer centricity, and ability to adapt to a complex and ever-evolving environment here​

Get comfortable for your 5 minutes of sparks to hack.

✨ Spark

AI expands expertise. And the same move creates friction in organizations.

✏ AI augments your expertise

Think of your T-shaped profile: deep expertise in one area (the vertical), and a broader reach across others (the horizontal).

AI deepens the vertical: more depth in your current role. AI widens the horizontal: you reach adjacent or borderline expertise. For me, this will be a legal review of contracts and online marketing best practices, or negotiations.

You can do things you couldn't do. You get faster to the data you need. You get an answer without waiting for someone.

That is good news. But when should we stop?

✏ The instant expert when too much is dangerous

The risk is becoming an instant expert: people mistaking access to information for understanding.

Search gave you sources to read. AI gives you a synthesis that feels like understanding. The pink area in the visual above is where excessive augmentation negatively impacts you and your organization.

This is "artificial certainty". A fluent, confident AI output makes uncertain things look knowable. In a meeting, the slide anchors the room.

I see it in my coaching. For 9 months, I have been coaching a 10-person market access leadership team. Lately, they complain about being challenged in business reviews, forecast meetings and strategy discussions by marketing or medical colleagues. The problem is not being challenged. That team's friction — it's good. The problem is that the marketing or medical person used poorly prompted AI sources that feel artificially certain.

✏ Impact on teams and collaboration

What does it cost a team? Expert authority erodes. Decisions get fragile. And trust drops when the error comes out.

So you end up with this chaotic organization ( see below).

So, is the instant expert the problem?

No. It is a symptom. The problem is that most teams have no shared agreement on when to cross a border, and who validates. This is one of the hot topics in the Coaching for Change for Leaders we runs, helping leaders in healthcare organizations to anticipate the issue of increased AI adoption and how to adapt their leadership style and processes to ensure high-performing teams (have a look here )

✨ Hack

Four hacks. Two for you, two with your team.

✏ Hack 1 : Make sure your AI says "I don't know"

Most AI tools are built to please. They agree, they cheer, and they rarely share data gaps or assumptions. So your colleague walks into the meeting overconfident. And sometimes, so do you.

As an organization, in your team and with your peers' teams: make sure people personalize their AI. Ask it to flag assumptions, show data gaps, and say "I don't know."

If everyone has an AI that is more honest, the overconfidence crossing borders goes down. (I shared my own setup in the last newsletter)

✏ Hack 2: Deal with the emotion, then use your go-to questions

Being challenged with a chatbot answer can feel personal. Notice the emotion first.

Then use your own framework: one or two go-to questions that help you get the conversation right. For example:

• "What did you ask the AI?"
• " Which sources did you use?"
• "If this is wrong, what would change in our decision?"

✏ Hack 3: Agree inside your team when not to use AI

Discuss as a team which borders you want to cross, which ones you don't, and how. If you work in marketing, how much compliance work do you want to do before sending documentation to the regulatory team?

Do you want to do the tasks of your lawyer or compliance lead? Or do you want AI to improve your questions and context, so you have better conversations with those experts? This is specifically the way I use it. I would prep every discussion with an expert, from home renovation to legal or health, with an improved list of questions and understanding of the main acronyms or terms.

✏ Hack 4: Create the conversation with your colleagues

Agree with the cross-functional teams you are part of how you use AI on expert topics. Say when content comes from AI. Define how you verify, when you consult the expert, and how you escalate when AI and expert disagree.

When it happens to several people from the same function (eg the market access example), ask what your team could offer its colleagues. For example, should market access build an Agent or Notebook with tips and advice that marketing uses, instead of a random ChatGPT chat without the right context? You know where the source comes from. Maybe you save time. It is context-dependent, but worth considering.

That is it for today.

See you in two weeks. Keep the spark alive!

Aurélie

What's up with me?

I enjoyed Halloween, and here at the Beyeler Foundation in Basel, the new garden and its giant spider from Louise Bourgeois.I also finished a coaching program for a global market access team and continue to do so, where I developed two new frameworks to help people see how AI could help them. I also assessed the Swiss Innovation Agency funding requests, a lot on AI and RD for pharma. It's quite an exciting field.  

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!