2 reasons why every CFO must be excited about AI Agents
Here’s my advice on AI (and what I’d tell my team):
Every finance person must be excited about AI Agents (2 reasons why).
1. AI Agents actually get things done.
ChatGPT Agent isn’t like Gemini, Grok, or Claude. It doesn’t just talk.
It logs into your systems, finds files, matches transactions, drafts emails, and prepares reports. Throw it messy ERP exports and it sorts them out. Feed it raw data, and it builds your first draft board deck or summary.
It acts like a junior analyst:
checks, thinks, acts, and double-checks. No hand-holding.
Here’s why everyone in finance must be excited about AI Agents:
Automate manual work. Agent logs into your systems, finds files, matches transactions, and drafts emails. It handles the busywork.
Turn chaos into order. Throw it messy exports from different ERPs. It lines them up and shows you what doesn’t match.
Prepares reports. Give it raw data. It builds the first draft of your board deck or financial summary.
Acts like a smart analyst. Agent checks, thinks, acts, and checks again. Without you babysitting every step.
2. AI Agents give you back your time.
You spend less time cleaning up files and more time on what matters. No more living in spreadsheets. No more chasing numbers or babysitting routine work.
But every AI agent relies on humans to perform well.
You can give it limited access. Want an agent to check your calendar only? Only share your calendar. Don’t open your whole inbox unless you have to.
Don’t set and forget. Review what Agent does. If it flags a mismatch, check it before you send.
Why and how to use it?
Why: You get your time back. Your team spends less time cleaning up files and more time on what matters. No more living in spreadsheets. No more chasing numbers.
How: I automated the entire process with ChatGPT Agent and honestly, I think this is what everyone in finance has been waiting for.
In this example below, I’ll walk you through how I used the ChatGPT Agent to automate the most boring finance task.
How to automate finance functions with ChatGPT Agent
When I talk to CFOs, they all worry about things like…
See what’s possible for your team.
OpenAI knows finance teams can’t afford careless automation.
That’s why they’ve built the strongest safeguards and user controls ever rolled out:
Permission-based access
Activity tracking
Robust logic checks at every step.
But, as with any new tech, it’s experimental, and no one can predict every edge case.
So here’s my advice (and what I’d tell my team):
Use Agent for automation you can supervise.
Don’t give it open access to sensitive emails or confidential files unless absolutely necessary
Always review actions and outputs. Especially for anything that affects financials, compliance, or reporting
Start small: let Agent handle repetitive, time-consuming workflows you can easily monitor
For example, it’s safe to have Agent match your transactions and prepare a reconciliation draft. But don’t let it loose on the entire accounting inbox with a fix everything command.
Those are the situations where something could slip through.
On Humanity’s Last Exam, an expert-level AI benchmark, the ChatGPT Agent’s model just hit a new SOTA (state of the art) score, 41.6 pass@1. And because the agent plans on its own and picks its tools, it can solve the same problem in multiple ways. With a simple rollout strategy. It tries up to 8 different approaches simultaneously and selects the best one. The agent’s HLE score jumps even higher, to 44.4.
Sam Altman just signed a deal for an additional 4.5 gigawatts of capacity with Oracle as part of the Stargate initiative.
It’s easy to throw around big numbers, but this is a gigantic infrastructure project.
Real-world proof of how fast AI is scaling up.
Stargate isn’t standing still, either. They’re already planning to expand beyond the $500 billion commitment announced back in January.
Progress photos from Abilene show the scale of what’s coming.
The world’s biggest players are doubling down.
AI infrastructure is growing at a pace we’ve never seen before.
The teams who learn to use these tools today will be the ones who set the pace for everyone else tomorrow.
Here’s what John, a finance operator, did with Agent:
John needed to extract hundreds of numbers from several PDF files. Agent accessed the files, pulled out about 200 figures from each PDF, and, on command, built a full spreadsheet, already formatted to his specs.
The result?
98% of the data was accurate right away. John fixed the rest in Excel with a simple copy-paste. He could have also made those tweaks directly in ChatGPT.
So If Agent can handle 90–95% of the heavy, time-consuming work, it’s a game changer. That’s hours saved on every project.
The Bottom Line
You don’t have to wait for some future version of AI. You can use agents today to take the grind out of budget work, reporting, and reconciliation, right where your business is today.
The teams who learn to use these tools carefully and deliberately will outpace those still stuck in manual mode.
This isn’t really about eliminating jobs. It never was.
It’s only about amplifying your best talent and building a smarter, faster finance function.
Try ChatGPT Agent on your low-stakes workflows.
Watch how much time you get back.
Meet the world where it is.
Move faster than everyone else.
That’s how modern CFOs win.
Whenever you're ready, there are 2 ways I can help you:
Promote your business to 21,000+ engaged CFOs and senior finance leaders by sponsoring this newsletter.
If you’re building an AI-powered CFO tech startup, I’d love to hear more and explore if it’s a fit for our investment portfolio.
I’m Wouter Born. A CFOTech investor, advisor, and founder of finstory.ai
Find me on LinkedIn