Jean-Charles Chemin is co-founder and president of Legapass, a Nice-based company created in 2021 to secure the transmission of digital and financial assets. An engineer by training who came through the French Tech Côte d’Azur ecosystem, he also chairs the Nice Start(s) Up association, which brings together startups in the region. Legapass initially positioned itself as a “digital vault” certified by the Conseil supérieur du notariat. But it pivoted to become the technological ally of notary offices. Thanks to solutions like Radar, it detects forgotten assets in estates, and VigiNot, which automates part of anti-money laundering (AML-CFT) obligations. Supported by the Adnexus fund, the investment vehicle of Groupe ADSN.
Our conversation helped understand how AI now permeates every layer of Legapass: from product design to development, from algorithm industrialization to strategic positioning in a highly regulated niche market.
From digital vault to estate radar for notaries
Initially, Legapass was born from a simple insight. The proliferation of digital accounts, crypto wallets, trading platforms and online credentials poses a major risk to estates. Part of the estate literally disappears into digital limbo. The startup’s first products therefore targeted the general public with a secure “vault” to organize the transmission of this sensitive information.
But the market would quickly redirect the trajectory. The first interested notaries raised a strong objection: the vault is useful, but too late in the value chain. They need an immediate solution to identify the deceased’s assets at the time of succession, not ten years earlier, depending on the adoption of a B2C service.
This is where the founding pivot occurs. The team begins to reuse, for the needs of notary offices, the “digital investigation” techniques originally imagined for the vault. Based on clues, identifiers, and digital traces, the algorithm can uncover hidden assets: trading accounts, crypto platforms, online contracts. One of the first emblematic cases concerns an octogenarian whose radar brings up “tens of thousands of euros” in crypto assets and trading, unknown to the family.
From these early successes, Legapass develops a platform interfaced with notaries’ professional tools, accessible via the ID.NOT ecosystem. The B2C vault will gradually be phased out to make way for the next generation of Legapass services. The core market clearly becomes B2B, in a very specific niche: valuation and securing assets in the context of estates.
Working daily with offices, Jean-Charles Chemin and his team detect a second, even more structural professional pain point. Notaries face a continuous increase in their due diligence obligations. Anti-money laundering and counter-terrorist financing, sanctions list verification, beneficial owner control, documentation and archiving of compliance evidence.
These tasks are critical. They are heavy. And they are very poorly equipped for small structures, which have neither the resources nor the culture of a banking compliance department. The European regulatory context (4th, 5th and 6th AML-CFT directives, Tracfin system, FATF recommendations) nevertheless makes a more industrialized approach to due diligence essential.
Building on the platform developed for Radar, Legapass adds a component: VigiNot. The solution draws in real time thousands of data from reliable sources (sanctions lists, FATF lists, beneficial owner registers, etc.), cross-references them with office files and generates time-stamped, traceable, auditable results.
The goal is not to replace the notary’s expertise. It is to provide them with a digital analyst, available 24 hours a day, who prepares the ground, pre-classifies risks, flags inconsistencies and documents proof of diligence. This is a typical case of AI applied to compliance: a massive volume of public and structured data, complex but codifiable rules, a strong requirement for traceability.
AI as a hyper-productivity engine for a small team
Behind the scenes, AI is first and foremost a speed lever for the development team. Jean-Charles Chemin talks about three levels of integration.
- Code copilots. Tools like GitHub Copilot that assist developers on the fly. They complete functions, suggest tests, accelerate the reading and rewriting of existing codebases.
- Code generation. Starting at Christmas 2024, the team massively exploits specialized models to generate entire blocks of application code. The productivity gain is such that they “jump on it” immediately.
- Ultra-fast prototyping. For each new product idea, the approach changes. Instead of spending days designing theoretical specifications, the team prototypes in a few hours. It tests several paths, compares results, discards what doesn’t hold up, keeps what works.
AI is no longer just an assistance tool. It becomes “a thinking ally.” It allows exploring varied paths, iterating quickly, accepting mistakes without irreparable cost. This way of working aligns with recent studies on “augmented” developers, which observe productivity gains of 20 to 50% depending on tasks, but also an increase in the perceived quality of delivered code.
The metaphor chosen by Jean-Charles is that of the transition from walking to horseback riding. AI allows going faster and further. But it also increases the possibility of falling. Hence a dual imperative: keep control of strategic direction and invest heavily in securing, industrializing and testing each AI component before scaling.
When AI also permeates the product
At Legapass, AI doesn’t just accelerate developers. It is at the heart of the features delivered to customers. In Radar, it helps detect assets scattered across banks, investment platforms, crypto wallets or digital contracts. In VigiNot, it sorts, classifies, matches, prioritizes masses of documents and regulatory information.
The idea is to offer notaries a “digital analyst” capable of working at a controlled cost while guaranteeing analysis quality comparable to that of a human expert. The company invests continuously and significantly in AI to improve its tools and, in return, increase the efficiency and value provided to notary offices.
This strategy resonates with a broader movement: European RegTech companies are increasingly relying on AI to automate KYC, AML-CFT, third-party screening, while finely documenting proof of compliance for regulators.
Digital sovereignty: an assumed pragmatism
Questioned about digital sovereignty and the choice of AI providers, Jean-Charles Chemin adopts a nuanced position.
On infrastructure and sensitive data, the line is clear: production servers are hosted in France, with French providers, outside American hyperscalers. Personal data and clients’ asset information do not leave this perimeter.
On AI models, the hierarchy is more pragmatic. The leader says he prefers a French solution, then European, before American or Chinese alternatives. But his priority remains the product and business: “I will not sacrifice my product solely to promote France or Europe.”
Concretely, the team regularly evaluates available models and does not hesitate to change suppliers if a European alternative offers comparable performance. Algorithms are designed to be interchangeable: what works today with one supplier can be switched tomorrow to another, for example to Mistral models for certain future products.
This very operational approach raises a broader question for European SMEs: how to reconcile the need to remain competitive with the desire to support a sovereign AI ecosystem? Can we make case-by-case decisions without losing strategic coherence?
Hypergrowth in a regulated niche
Legapass remains a small team. About twenty people, including apprentices and interns. But in its target market, the startup claims a dominant position. Approximately 45% of the French notary market for its key solutions, tens of thousands of reports generated daily, sometimes hourly, and continuous growth since the beginning of the year.
This hypergrowth poses structural challenges. The team has managed to absorb the increase in activity without massive hiring in operational support, thanks to AI and strong automation. But to innovate and create new products, hiring will be necessary.
The dilemma is classic for profitable B2B startups. The company is self-financed, has repaid its debts, generates profits. It does not need to raise funds to survive. But to address new markets—other regulated professions, foreign notaries, new use cases within the notary profession itself—it will need to structure, recruit, accept a certain organizational complexity.
Jean-Charles Chemin is cautious. He knows the risks of too rapid growth: middle management layer, heavier processes, loss of agility, dilution of product culture. His approach is based on three axes:
- Preserve profitability and cash flow. The company has already achieved this first objective.
- Identify a “second flagship product.” Either for French notaries, or for a new regulated profession (accountants, auditors, lawyers, bailiffs), or for notaries in other European countries.
- Only diversify where “the right person” is found. Rather than imposing top-down development, the leader first looks for individuals capable of carrying a market: a local contact in Belgium, a business expert on the accountant side, etc.
Humans remain at the center. Artificial intelligence is everywhere in the processes, but the critical variable remains the human intelligence capable of understanding a market, a profession, a regulation.
AI-first culture and selective recruitment
This vision is reflected in the recruitment policy. Legapass is not just looking for developers or business profiles. The company wants collaborators capable of mastering AI, understanding its levers, measuring its risks.
Generative AI is used to prototype, think, document, automate. But it should not be used to follow fashion or to “do AI for AI’s sake.” Use cases are carefully selected. Where a simple classic algorithm suffices, the startup does not outbid by adding an unnecessary AI layer.
This positioning joins the current debate on “AI-washing,” where some players highlight AI essentially for marketing purposes, without real functional added value. Legapass claims the opposite: AI as a discreet but decisive tool, serving a concrete problem.
A clear message to SMEs: the real risk is falling behind
At the end of the conversation, I ask Jean-Charles what advice he would give to SME leaders who are still hesitant to embark on an AI approach. His answer is unambiguous.
It’s not about “jumping everywhere” or piling up trendy tools. It’s about recognizing that we’re experiencing a technological revolution. And that each lost day creates a delay that’s difficult to catch up. According to him, the only truly critical risk is that of inaction.
In practice, this involves several steps:
- Identify a few time-consuming, repetitive, documentary processes where AI can produce quick gains.
- Train, or train an internal core, in responsible uses of AI (data quality, confidentiality, security, intellectual property).
- Prototype quickly, test in real conditions, measure gains, document risks.
- Decide what should be internalized, what can be entrusted to specialized suppliers, what should remain outside the AI perimeter.
