How We Built an AI Agent for Our Own Finance Team

MoneyPenny, our AI Agent, now manages our bank account. 💰🤖
We walk our talk.

We’ve been adding new skills to our AI Agent, and we decided to put them to work in our own Finance function.

Our bank is Revolut Business (we love them), and here’s what MoneyPenny can now do:
🏦 Daily Bank Reconciliation
→ Update accounting
→ Detect and escalate exceptions
→ Collect missing invoices and prepare them for accounting
💰 Accounts Receivable
→ Detect incoming payments
→ Match payments with invoices
→ Track paid/unpaid invoices
→ Eventually, send payment reminders
💳 Accounts Payable
→ Monitor payments
→ Match payments to invoices
→ Flag discrepancies
📊 Month-End Closing
→ Reconcile accounts
→ Check for missing invoices
→ Prepare reporting

And honestly… it feels pretty good to be doing this ourselves.
Most of our projects are currently on hold because, well… you guys are all on holiday 🙄☀️

So we’re using the time to:
→ Reuse code and processes we’ve already built for clients
→ Test them in our own Finance function
→ Prepare for a client with a very similar setup

And, most importantly, we’re building it the way we believe AI Agents should be built:
🔐 Secure: no unnecessary access or risk to the bank account
🎯 Reliable: clear rules around what the Agent can and cannot do
🛠️ Owned: no dependency on another external tool or licence
⚡ Impactful: the Agent actually does the work

If you have a similar Finance process and want to know how we built it, comment or DM me.

Enjoy the summer! ☀️ We definitely are with MoneyPenny. 🤖
Alexis

—

Originally posted on LinkedIn:

https://lnkd.in/p/d2-XTV25