Ember.Ember
All case studies

Second Brain

Le Labo des CEO tests its strategic analyses against several AI models with Ember Second Brain

Portrait of Mélanie Louchart

Mélanie Louchart

CEO, Le Labo des CEO

LinkedIn

2

different recommendations for the same question, with ChatGPT and Claude

6

steps in her way of working with Ember Second Brain

4

separate apps, at least, where her analyses and documents were spread

Mélanie Louchart supports business leaders with their strategic decisions. Before Ember Second Brain, asking ChatGPT and Claude the same question gave her two different recommendations. She now works with an AI assistant for business leaders: a single space where several artificial intelligence models challenge each other, with the company's context.

One question, several lines of reasoning in return

The first use Mélanie Louchart describes comes down to one action. She submits a question to Council mode, and several models examine the same problem. They weigh their reasoning against each other, then bring out the points of agreement, the disagreements and the questions to dig into.

The models also take into account the context available in the platform. She no longer has to copy the same question from one app to another to compare the answers by hand.

It is the feature she uses most for important topics.

Two well-written answers, and a decision to make

Mélanie runs Le Labo des CEO, a company that supports business leaders on their markets, business models and funding decisions. She had been using ChatGPT and Claude for a while. These tools brought her real value for researching, organizing and writing.

The problem showed up on strategic topics. She could ask both tools exactly the same question and get two different recommendations. Sometimes both answers seemed perfectly argued. Nothing in either one made it possible to say which was right.

So she had to check, ask again and cross-check. That took time. With each new conversation, she also had to explain the context again: the business, the positioning, the clients, the business model, the goals and the constraints.

These differences are understandable. Models do not always use the same reasoning methods, or the same knowledge or assumptions. For a leader who has to decide, the difficulty remains: how do you choose between two analyses that both seem credible?

Six steps, from topic to document, in a single space

Her process has six steps. She starts from a concrete question, such as a market, a business model or a marketing strategy, and sets out the context and the goal. She then calls on a specialized skill from the library, which covers strategy, marketing, finance or organization.

When the topic is a big one, she turns on Council mode. She reviews the arguments and the points to watch, then digs into the most important questions. Once her thinking is structured, she asks to generate documents with AI: analysis reports, Word documents, Excel spreadsheets or action plans. She does not need to switch apps to go from thinking to a document.

Illustrative example: imagine a business leader who wants to sell a new service. One line of reasoning might favor speed to test demand, another might highlight the risk of a vague positioning, a third might stress profitability. Council mode then brings out the trade-offs: launch right away, run a limited sales test, revise the price or strengthen the promise before investing more.

For topics outside her field, such as a legal, tax or HR question, she can be pointed to expert partners. Artificial intelligence prepares the file, then a professional takes over.

A memory that keeps the context from one conversation to the next

The company's information stays available for later analyses. Mélanie does not have to present her whole business again with every question. The answers are better contextualized, and the analyses gain in continuity.

Before, her documents were spread across several environments. An analysis in ChatGPT, a business plan reviewed with Claude, spreadsheets in Excel, documents in Word, visuals in other apps. She describes this scattering as a complication for the continuity of her work.

What changes day to day, and what remains to be measured

Mélanie says she feels more at ease with the recommendations she gets. She keeps her own judgment. She also knows that several models can share the same errors or rely on the same incomplete information. Weighing them against each other helps spot disagreements and assumptions to check. It does not guarantee the right answer.

What changes most is continuity. When she works on a topic related to the business, she uses Ember Second Brain. She keeps using ChatGPT and Claude when they suit her. She also has access to specialized skills to deepen market analyses and business plans.

For a market, she is not just looking for a list of competitors. She wants to understand customer needs, industry dynamics, opportunities and risks. For a business plan, she questions the revenue and cost assumptions, and asks whether the target customers are defined precisely enough and whether the forecasts hold together.

She does not put a number on the time saved at this stage. What she describes is a way of working: a question, a debate between models, a document, then an expert if needed.

Frequently asked questions

Your next decision can start here.

Describe your priority. Ember helps you move forward.