With artificial intelligence (AI), finance will face a labor shortage, contrary to received wisdom, argues Vincent Aurez, AI and finance expert, president & co-founder of FIGEN AI, columnist for Capital.fr.
Vincent Aurez
At the beginning of the 19th century, 65% of French workers were employed in agriculture. In 1936, the primary sector still accounted for 32%. Today, agriculture employs only 2.3% of the workforce, or 681,000 people.
Does that mean we have 30% or 60% unemployment? No. The unemployment rate stands at 7.7%, or 2.4 million people. It's a lot, certainly, but far from the 20 million that such logic would imply.
Where did the others go? They shifted to jobs that didn't exist before. Between 2021 and 2023, 1.3 million French people worked in a digital occupation, or 4.6% of the workforce versus 2.7% in 2009. According to INSEE, the "data analysis and AI" family is already growing within these occupations.
I hear the same worry everywhere: AI will destroy jobs in financial advisory. Wealth management advisors will disappear. Analysts are doomed. What if it's the opposite?
— Vincent Aurez
Jonathan Ross, founder of AI company Groq acquired by Nvidia, does not foresee a job shortage because of AI. He instead anticipates a labor shortage, with too many unfilled job openings. The distinction deserves attention.
With AI, new markets open
A McKinsey study estimates that 63% of tasks in financial advisory could be automated. A wealth management advisor currently spends between 40 and 65% of their time on administrative tasks: compliance, reporting, data entry. AI can absorb that work. The question then becomes: what does the advisor do with that freed time? They advise.
The problem for finance will not be an excess of advisors; we are already lacking them. According to the AMF, 80% of CGPs fail to achieve full compliance, not due to incompetence but due to lack of time.
Regulatory complexity is exploding: MiFID II regulation alone amounts to 1,200 pages. A dedicated AI solution reverses this equation.
Let's go further. Ross's argument goes beyond simple reallocation of existing time. If AI reduces the production cost of financial services, prices fall, demand increases and new markets open.
Potential demand is massive. Early 2024, 90.5% of households hold financial assets. Yet this wealth remains highly concentrated: half of the wealthiest households hold 92% of assets. The broader finance sector accounts for 1.36 million jobs, or 4.7% of total employment. It's not marginal, but it's not commensurate with the need.
With AI assistance, one advisor can support three times more clients. Millions of underserved households become reachable. We questioned job scarcity at the start of this piece; now we're measuring market creation potential.
And then there are jobs that don't exist yet. In 2000, no one hired community managers or SEO strategists. In 2010, the data scientist role was niche. In 2020, few knew what a prompt engineer was. The jobs of 2035 are as foreign to us as a web developer was to a 1925 farmer.
Employment: emerging new jobs in finance
In finance, I already see a few emerging: architect of wealth agents, auditor of financial algorithms, AI transition coach, manager of autonomous agents. I don't know which of these jobs will truly exist—who does?
The reasoning error is always the same. We see the jobs that disappear because they already exist. We don't see those that will appear because they don't yet exist. Losses are visible and immediate; gains are future. A financial advisor could explain this to you better than I can.
Some jobs will disappear. With AI, repetitive low-value tasks will be automated. But the advisor who spends time with clients, understands their anxieties, and supports them through hard decisions has never been more useful.
Key points
- "With AI, repetitive low-value tasks will be automated."
- "AI does not replace judgment. It does not replace empathy."
- "With AI assistance, an advisor can support three times more clients."
AI does not replace judgment. It does not replace empathy. It does not replace the ability to look a client in the eye and tell them no, this is certainly not the right time.
In five years, we will look at this period as we look at the 1990s today, that is, 35 years back. It is this acceleration of technological and societal change that can sometimes be dizzying.
