AI in Finance: From Buzz to Business Value, with Nicolas Boucher
Summary
In this episode of The Growth-Minded CFO, co-hosts Lauren Pearl and Alex Louisy sit down with Nicolas Boucher, AI for Finance Expert and Trainer.
Nicolas Boucher: From early adopter to finance AI educator
In late 2022, when ChatGPT first launched, much of the buzz was focused on gimmicky entertainment and marketing use cases. Nicolas Boucher had a different reaction.
As an experienced finance professional, he immediately saw potential beyond poems and press releases. He began experimenting with how AI could transform workflows in corporate finance—use cases no one else was talking about.
Nicolas's curiosity led to his first guide, How to Use ChatGPT in Finance, which quickly reached more than 5,000 downloads. It marked the start of his journey as a leading voice on the intersection of AI and finance, followed today by over 1 million people on LinkedIn.
The moment everything changed: from text to calculations
For much of Chat-GPT's early public rollout, one major limitation held it back in finance: it couldn’t calculate reliably. Generative models were probabilistic—they gave answers that sounded right, but often weren’t.
That changed in mid-2023.
Nicolas recalls the first time he uploaded a file into OpenAI’s Code Interpreter (now called Advanced Data Analysis). It analyzed the data, built a heat map in Python, and even offered insights—tasks that Excel couldn’t handle alone. It marked a shift: AI could now help with real quantitative work, not just writing.
"I remember storming out of my room, going to my wife and saying: 'Now this is really going to change a lot of jobs because we're moving from writing to calculating'."
On hype, headlines, and real product value
Many finance leaders are still trying to separate what’s real from what’s marketing when it comes to AI.
The Growth-Minded CFO Podcast co-host Alex Louisy reflects in this episode how he saw firsthand how products rushed to add “AI-powered” to their branding in 2023. In many cases, those tools did little more than write basic emails. The value was superficial.
Nicolas offers a practical lens: don’t ask if a tool is powered by AI—ask if it solves a real finance problem well. He points to Trullion as a good example: the platform is built specifically to help with revenue recognition, parsing contracts to extract compliance requirements under US GAAP and IFRS.
It’s not flashy. But it works—for the right company with the right problem.
AI won’t replace you, but misusing it might
One theme that comes up again and again in the episode: using AI the wrong way can backfire.
Nicolas sees some finance professionals handing off work to AI tools they don’t fully understand. Then, when the result is wrong, they blame the AI.
“It looks so intelligent that you might want to take a shortcut,” he says. “But if you present something wrong to your boss and say, ‘It was the AI,’ your boss might just say, ‘Then I’ll just work with the AI instead of you.’”
His advice: use AI to sharpen your analysis, not replace it.
Why context—and not the model—matters most
Most large language models produce similar output. The difference, Nicolas says, is how well they fit into your existing workflow.
Take Microsoft Copilot as an example. If your company already operates within the Microsoft ecosystem, Copilot can tap into your emails, meeting notes, and files to give answers that are deeply contextual. Instead of asking ChatGPT for a project charter and explaining everything manually, Copilot can piece together the right content from past conversations and documents.
That’s the real differentiator: context.
Leading AI like you lead a team
Nicolas encourages CFOs to think of AI as another team member:
Recruit the right tool
Instruct it with clear prompts
Review its work
Train or replace when needed
It’s not plug-and-play. It’s ongoing management—just like with people.
And for finance teams willing to embrace that mindset, the upside is real.
Key takeaways for finance leaders
The best AI tools today are narrow, not general-purpose—they solve specific problems very well.
AI can now assist in real analysis, not just writing, thanks to advances like code interpreters.
Misuse or over-reliance on AI risks credibility—understand the output before you share it.
Integrations and environment (like Microsoft Copilot if you're a company running on Microsoft) matter more than model performance.
CFOs will need to learn how to “manage” AI tools just like they manage people.
Full Episode
Hear more from Nicolas Boucher by tuning in to the full episode:
Listen (and subscribe) on Spotify