Hey Finance Engineers,
I feel like every finance team right now is living with what I like to call the keyhole problem.
We’re all pressed up against the door, trying to catch a glimpse of what the others are doing with AI.
And all anyone gets through that tiny keyhole is a Tesla flying car. A vendor’s post on LinkedIn. A line from a conference stage. A podcast where someone says their finance function is “AI-native.”
Everyone has heard of the AI-native finance function, even if nobody quite agrees on what it means. But very, very few people have had a real, insider look at how one actually runs.
How does a finance team go from using ChatGPT for emails a year ago to running on agents today? What was step one? And what are these mystical “agents” everyone keeps mentioning, and which processes do they actually cover?
I’m done with keyholes and flying cars in finance.
So in this issue, I’m opening the door on a real enterprise finance team that went AI-native.
You all know Chargebee as the software company that builds billing, metering and monetization infrastructure, I’m sure. I do. But their own finance team’s AI journey is a story nobody has told yet.
To get it, I sat down with Neerav Patwa, Finance Director at Chargebee. He walked me through the whole year of their AI transformation: the first script that failed, the aha moments, and the 50+ agents and AI automations the team runs today.
Inside, you’ll get the path, the rules they set along the way, a dozen of those agents, and a deep dive into two of them.
Let’s go.
Chargebee’s path to an AI-native finance team
A year ago, Chargebee’s finance team was where most of us were.
ChatGPT, used for emails. Meanwhile, the rumors were getting louder. Cursor was booming. People with no engineering background were posting working code. And finance teams were wondering: could we do that?
Chargebee had already set the stage. They gave every team generous, org-wide access to AI tools, not just engineering and sales, because they believed inventive automation ideas could come from any function. Finance included.
So one person on the close team decided to find out. He picked the task everyone hated. Capitalized software accounting looked like this: 16 director spreadsheets, consolidated by hand every quarter. Six to seven days, two people. He asked ChatGPT to write an Apps Script. It failed. Two months and 25 versions later, it ran end to end.
Seven days across two people became a day’s work.
A finance person had just killed a six-day process by himself. If he could, anyone could.
“It was an aha moment for the entire finance team: such an extensive process could be automated by approaching it with first principles and AI.”
— Neerav Patwa, Finance Director at Chargebee
So the path to an AI-native team began.
