Back to the blog

·6 min read

Series “A day with AI agents” · Part 2

A day in finance with AI agents

The director wants to know why margin dropped and what to expect this quarter. Follow a finance analyst who answers with the number and where it came from, without opening five spreadsheets.

In finance, almost every question sounds simple and almost no answer is. "Why did margin drop?" means matching the ERP, two spreadsheets, a contract and a conversation someone had with a supplier.

Rafael is a finance analyst at a company that uses Kortia. He's an illustrative character, but his day is made of situations every finance team knows. This is the second piece in the "A day with AI agents" series.

8am. What went off plan

Rafael starts the day with a question to the brain: "what went off plan in this week's expenses?". The answer matches the ERP entries against the budget and points to where the gap is.

The question doesn't change; the time does. Instead of opening every report, he starts by looking at what matters, and every figure in the answer comes with its source: the entry, the spreadsheet, the contract.

10am. The director asks

A message arrives from the director: "why did margin drop this quarter?". That question used to cost a whole afternoon of spreadsheets.

Rafael asks the brain and finds the thread: two supplier contracts were adjusted, and the increase hasn't been passed on to clients yet. The brain shows the entries, the contracts and the meeting where the adjustment was agreed.

He checks it, adds what he knows about the business, and answers the director the same day, with the number and the explanation.

11am. The budget meeting doesn't get lost

In the budget meeting with the other areas, Rafael doesn't need to write everything down. The conversation is transcribed, and the brain drafts what comes up: a decision to hold back an investment, a new travel limit, a question from sales.

He approves what's right. The next steps become tasks with an owner, linked to the meeting where they were agreed.

2pm. The quarter's projection

Time to look ahead. Rafael asks the agent for a draft projection for the quarter, based on history and the contracts in force. The draft arrives with its assumptions in plain sight: what was considered and where each number came from.

He adjusts the assumptions with what he knows and what the market suggests. The agent builds the base; the projection that goes to the board is his.

5pm. The month's summary without building it by hand

At the end of the day, the director asks for a summary of the month for tomorrow's meeting. Rafael asks the agent for a first draft with the main figures, the deviations and the explanations that came up during the day.

What used to take hours of copy and paste becomes a review. He checks it, cuts what isn't needed and sends it.

What changed in Rafael's day

The difference isn't one big automation; it's the time that went back into analysis:

  • Questions answered with the number and its source, without hunting through spreadsheets.
  • Deviations explained by their cause, not just their size.
  • Meeting decisions recorded and tied to tasks.
  • Projections and summaries that start from a draft, not a blank page.

The agent doesn't decide for finance. It clears away what used to eat Rafael's time: searching, matching and assembling. The decision stays with whoever knows the business, now with the data in hand.

In the next piece of the "A day with AI agents" series, we follow a day in marketing.

WhatsApp

Let's talk on WhatsApp

Tell us who you are and we'll reply there.

You send the message from your own WhatsApp. No commitment.