Wealth professionals are increasingly demanding regarding the software solutions offered to them. Vincent Aurez, formerly of Novaxia, is one of the partners of the Fintech Figen AI, which has just launched a new complementary tool developed using artificial intelligence.
How can Figen AI be defined?
Figen AI is a suite of intelligent assistants designed to support wealth advisory professionals in their daily work. Concretely, our first product launched in January is an expertise search assistant capable of answering any question related to financial or extra-financial knowledge.
Before Figen AI, the advisor had to consult encyclopedias manually, in paper or digital form, which could take several hours. Today, thanks to AI and our technology, they obtain an answer in one minute, enriched with all the sources they would have had to consult themselves.
What is your added value in this competitive software market?
Wealth professionals are not simply looking for technical answers. They need a tool that fits into their way of thinking, into how they support a client. That is why Figen AI is much more than a search engine: it is a specialized tool, a true work companion.
Today, more than 530 firms use Figen AI in their professional daily routine.
Does your tool compete with Harvest, which is currently going through a difficult period?
Our first assistant was launched in January, even before Harvest's situation deteriorated, with a strong conviction: historic players like Harvest were built on a technological revolution, that of digital and the Internet. They were pioneers in their time. But the arrival of generative AI, notably ChatGPT, marked a new technological break. We saw a clear opportunity because no player at the time was natively developing tools truly designed for this sector in the AI era. That pushed us to create Figen AI. So we can be seen as a complement to Harvest.
We also have a major asset: our development costs are much lower, which allows us to offer ultra-competitive prices. We notably equip young firms today, which previously couldn't afford wealth solutions billed at several thousand euros, as well as institutional players (groups and private banks). And above all, we innovate where others do not go: for example, no tool in France automatically prepares suitability reports.
How to extract the essence of artificial intelligence without falling into its pitfalls?
AI serves the profession, not the other way around. Artificial intelligence has a remarkable capability, but also a major flaw: what is called hallucination, i.e., its tendency to invent answers.
To solve this problem, we built a domain-specific AI. How? By creating, with a team of wealth engineers, a proprietary database, constantly updated, which becomes the AI's sole source of information. Result: the AI only draws from this database—no random searches on the Internet—which reduces hallucinations to zero and guarantees reliable, verifiable, and contextualized answers.
Next, our expertise assistants rely on databases far larger than those of traditional tools:
- For everything related to wealth or tax expertise, the AI relies exclusively on our proprietary content.
- For other cases, such as fund or SCPI analysis, we enrich the analysis with regulatory data (prospectuses, DIC), supplemented by structured monitoring on the Internet.
But again, the AI always indicates the sources used. Our client users therefore know exactly what each answer is based on: either a certified tax truth, or a reasoned and transparent monitoring. And to top it off, we have signed exploitation licenses with specialized data providers, notably on themes such as biodiversity and sustainable finance. For example, we have a partnership with the Caisse des Dépôts, which allows us to integrate and cite their reference studies. It's more relevant, more rigorous, and above all always up to date.
Our users tell us they gain up to one workday per week. And that's the whole strength of AI: not to replace the human, but to allow them to refocus on what matters—the client relationship.
Who are your potential clients?
- Young wealth management firms They are often our first users and several hundred adopted it at launch. Why? Because they do not yet have the means to have a full internal team. Figen AI then becomes their operational support, saves them precious time, and allows them to offer a level of service equivalent to much larger structures.
- Students in wealth management They are already familiar with AI, notably ChatGPT. Very quickly, they perceive the potential of a specialized tool that speaks their professional language and answers pertinently where generalist AIs reach their limits.
- Independent wealth advisors (CGPs) Used to using several often costly tools, they find in Figen AI a single platform, simpler, faster, and above all more economical, which centralizes their needs: expertise, simulation, compliance, etc., and integrates via API with wealth aggregation tools.
- Wealth engineers and multi-family offices (MFOs) Some of our assistants target very specific uses, such as complex financial simulations or inheritance devolutions.
We also see growing interest from legal professions, such as notaries or tax lawyers, who use the tool for quick, documented research on specific topics.
Are you referenced by CGP associations?
We chose to first prove our usefulness in the field before seeking to be referenced by CGP associations. We wanted to first prove that our solution met real needs. And that was the case: with more than 530 client firms, the proof is made. Today, we are in active discussions with all the major associations, which are currently testing the platform and giving us feedback. The objective is clear: to be referenced, but on the basis of recognition of the concrete value brought to professionals. Especially since our tool can be even more compliant than others, because it is configurable according to each structure's internal rules. For example, we have a "wealth assessment" assistant perfectly aligned with regulatory requirements.
However, we made a strategic choice: not to develop wealth aggregation ourselves, because it is a profession in its own right. Instead, we form partnerships with existing aggregators. The first we announced is Wealthcome, with whom an integration is being deployed.
There are a multitude of AIs available, what differentiates them?
Under the hood of Figen AI, there are not one but 17 specialized artificial intelligences working together. Each of them has a precise role:
- an AI expert in macroeconomics,
- an AI dedicated to SCPIs,
- another specialized in fund analysis,
- yet another in comparing financial products… and so on.
They cooperate with each other to provide an expert, precise, and contextualized response.
Artificial intelligence is entering a new phase of structuring: the first brick called "generalist AI" is composed of large models like ChatGPT, Mistral, LLaMA or Gemini. These are powerful foundations, but not specialized.
The second brick called domain-specific AI corresponds to the application layer. This is where Figen AI comes in. We select the right models, according to their respective strengths, for each domain assistant.
We are technologically agnostic! In other words, we use all available models without tying ourselves to a single provider, to choose what works best for each use. This pragmatism allows us to offer maximum performance, with total adaptability to sector evolutions.
“Take the example of a life insurance contract. The insurer knows perfectly the specifics of its own contract, but a wealth advisor needs an intelligent assistant that understands both life insurance taxation in general and the particular subtleties of each contract, notably regarding specific clauses. They do not want a generic answer about life insurance, but for example: which clause applies in the event of the first spouse's death, in such specific contract? And since no contract exactly resembles another, our "life insurance assistant" solution fully meets this need; moreover, it is updated with their contractual documents, which guarantees maximum precision, up-to-date and compliant.”
