La Story (Les Echos) Analysis: AI as an Investor's Compass—The 'Truth About Fees' Shaking Up the Asset Advisory Industry

▶ Original episode: https://shows.acast.com/la-story/episodes/lia-boussole-des-investisseurs-ils-se-confient-a-lia-comme-i

July 25, 2026

Overview

This article provides an analysis of an episode of "Les Echos de l'IA," a spin-off of the podcast "La Story" from the French economic newspaper Les Echos. Journalist Samir Touzani interviews Guillaume Fontenot, an asset management advisor (known in France as a Conseiller en Gestion de Patrimoine, or CGP) who has also developed his own AI assistant, about the new habit rapidly spreading among individual investors: "consulting AI before making investment decisions."

According to the latest survey by the AMF (Autorité des Marchés Financiers), the French financial market regulator, 11% of French people already use AI before making investments, and that figure rises to nearly one in five among those under 35. Fontenot feels this number is actually an underestimate, testifying that the vast majority of his own clients are using AI.

What is interesting is that AI is described not just as an information retrieval tool, but as a force capable of shaking the very structure of the asset management industry. Fontenot points out that by exposing the "weight of fees" and the "deception of sales pitches," AI will deal a fatal blow to the traditional business model of the advisory industry.

Ultimately, the outlook presented is that as knowledge becomes "commoditized" due to AI, the advisor's role will transform from technical consultant to "life design coach" and "asset psychological counselor."

Investors who 'confide' in AI

Fontenot first emphasizes the "willingness to confide" that investors show toward AI. He is surprised by how his clients confide things to AI that they would never tell an asset advisor.

The reason, he says, is that there is "no wall" between them and the AI. People talk to AI about extremely private topics, such as their children or arguments with their spouse over the weekend. Fontenot believes this is the true power of AI.

Because AI responds based on context, it becomes increasingly personalized as it accumulates the user's life through daily questions. He describes this as a "second brain," expressing his intuition that an AI that has absorbed every aspect of our lives can become an excellent asset advisor.

What AI is good at, and what it is not

Fontenot explains that investors get the most value out of consulting AI when asking fact-based questions. AI answers textbook-style questions—such as what a life insurance contract is, what a PEA (Plan d'Épargne en Actions, a French tax-advantaged account) is, or what stocks are—very accurately.

As a concrete example, he cites the 2025 income tax return. He says that many taxpayers were able to save a vast amount of time with the support of an AI that had been fed the tax authority's practical handbook.

On the other hand, AI is not good at "predicting the future." According to Fontenot, AI is an excellent theorist but extremely poor at future forecasting. The logic is that since no one knows the future, there is no reason to expect AI to be better at predicting it than a human. While it is good at explaining theories, its ability is limited when it comes to actually choosing which product to buy.

The pitfall of internet 'bias'

As another limitation, Fontenot cites the "bias" that AI carries. He points out that while AI feeds on information from the internet, the influence of bias on the web is very strong.

For example, if you ask an AI "what is the best life insurance contract," it will search for life insurance rankings on the web. However, these types of rankings are often heavily distorted by the commercial interests of affiliated companies. Fontenot warns that one should be very careful with such rankings, and by extension, the AI answers that cite them.

He also admits that there are still some issues regarding so-called "hallucinations." In particular, Fontenot has observed that AI hallucinates when dealing with numbers and standards, such as income tax tables. He says AI struggles to distinguish between tax tables for 2024, 2025, and 2026.

However, he also states that this problem is solvable. If you personalize the AI and provide reference materials like practical handbooks, instructing it to "follow these standards," the hallucinations stop. Therefore, he emphasizes that a "double check," where one always asks for sources and verifies the content by returning to the original information, is essential. He adds, however, that this is the same as how one should always verify an advisor's advice, as no one is infallible.

AI 'Does Not Lie'—A Threat to the Industry

Fontenot says that when it comes to subjective judgments like 'which investment should I choose' or 'should I invest in Bitcoin,' AI does not provide financial advice and therefore lacks that capability. In other words, while rational and factual opinions are requested, subjective opinions are not.

However, the point he emphasizes most strongly here is the nature of AI as something that 'does not lie.' Even if AI can hallucinate, it speaks the truth. And that truth, he says, is a truth that sometimes troubles the industry.

Specifically, AI clearly demonstrates the impact of fees on investment performance, and as a result, it is excellent at recommending ETFs (Exchange Traded Funds). Fontenot points out that this is exactly what is dangerous for advisors. Because when AI explains to an investor that 'thanks to ETFs, your performance will be better,' it is equivalent to implicitly saying that 'wealth management advice is too expensive.'

Fontenot states, 'This is why advisors fear AI,' adding that AI dismantles commercial sales pitches. Using the French expression 'empapaouter' (to trick or hoodwink), he concludes that while investors will no longer be deceived by sales pitches, the industry will no longer be able to string together nonsense to hoodwink clients—and that is the 'revolution.'

A Self-Made AI Assistant—'My Clone'

Fontenot is not just a user of such tools; he is also developing his own AI-based assistant. Using the unique phrasing, 'I like to cut off the branch I am sitting on,' he says that if it is a tool that will change the market anyway, it is better to have fun making it yourself.

He has built eight or nine theme-based AI agents based on ChatGPT. These include ones that assist with tax returns, analyze the real estate market, and help prepare for real estate investments.

As for what the added value is, Fontenot explains that he has implemented the experience he has accumulated as a CGP (Certified Wealth Management Advisor), the materials he has gathered, his own unique work methods, and his way of thinking into the AI. The goal was to make the AI 'think like me.'

As a result, he says that clients sometimes tell him, 'It's like talking to your clone.' He has even set the AI to ask clients about his favorite concept, the 'life project' (projet de vie)—that is, his strong insistence on 'understanding why you are doing it' and that 'money is energy that serves life.'

How Will the Advisor's Job Change?

Since AI can do all of this, is it a threat? Fontenot's analysis is clear. Current advisors earn their compensation from commissions on the products they sell or retrocessions (back-end commissions), not from the advice itself. He says that what AI is truly challenging in the industry is this 'marketing veneer.'

Whether depositing funds with an insurance company or opening a PEA to put ETFs in, financial institutions are still necessary. In other words, 'salespeople' who play a commercial role remain indispensable. On the other hand, he says that those at risk are the advisors who have justified the added value of their advice through their own expertise.

In the past, CGP expertise was rare and therefore deserved high compensation. However, now that knowledge has been 'commodified' by AI and is no longer rare but rather excessively abundant, that expertise can no longer justify high compensation. Fontenot points out that this is where the disruptive change lies.

In his view, because AI answers 90% of investors' questions, the wealth management industry will split into two. One is the 'commercialization industry' that only sells contracts, which accounts for the majority of the industry. The other is a very small segment where the situation is so complex that an experienced human is still needed.

Becoming a 'Wealth Psychologist'

Finally, when asked how the relationship with clients will change amidst these drastic shifts, Fontenot says that he is gradually becoming a 'life project coach' and a 'wealth psychologist' (psychologue patrimonial).

He says that the content of conversations with clients has shifted to children, family, divorce, and life ambitions, and they no longer talk about financial products. The topic has moved to 'how to use your money to build a life, realize projects, and be happier.'

Fontenot, who was originally a 'super-technician' whose weapon was purely asset technology, reflects that now, with more experience, he still has that technical ability, but he no longer puts it at the forefront. This is because the true added value lies in the ability to understand people, understand families, and understand projects, and technology serves in the background. He concludes with the words, 'The most difficult thing is not the technology, but the lives of families.'

#France #Economy #LesEchos #OpenAI #ChatGPT #GenerativeAI #AI #AssetManagement #Investment #ETF #Fintech #FinancialAdvisor #FrenchEconomy