This CFO went from 60 Hours to 20.

Mark’s a CFO in my community.

He went from working 60 hour weeks, to 20 using AI. And used the time he freed up to take on 5 new client projects and increase his income.

So next Tuesday at 5PM CET, I'm running my free masterclass to show you can can do this (and more). I'll show you:

• How to build a full bank reconciliation with AI while you still make your Friday plans
• How to build a 5-year, 3-scenario forecast in Excel in under 10 minutes
• The one Copilot setting that turns a basic, ugly dashboard into something you can present to your board

1,107 finance pros joined me last time. There are not many other ways to get 60-mins of my time for free.

So, save your seat here - and start putting hours saved towards your next big career move ;)

Super quick before we start - this is the first newsletter in a series on AI ROI. If you enjoy it, reply and tell me where you want to see ROI from AI in your Finance Team, I will answer in upcoming editions.

The lazy intelligent person always gets more work

When I was young, I read a book called "Réussir quand on est paresseux" - or - "Succeeding When You're Lazy." (by Ernie J. Zelinski)

The idea is simple. A lazy person hates a task, so they finish it fast. They do in 8 hours what the person next to them needs 12 hours for.

And what do they get for it?

More work!

I told a former CFO the same thing on a 1:1 call. She is a chartered accountant who has spent the last 8 years on finance transformation for big companies.

I told her, if you’re not careful, AI does not give you free time. It opens holes in your week, and you fill them by doing more.

I predicted this way before people started noticing. Aruna Ranganathan at Harvard Business Review followed a US tech company of about 200 people for 8 months. Nobody was forced to use AI. But the ones that did took on wider jobs that took more hours of the day. Jobs that nobody had asked them to do.

But how do you capture those productivity gains?

Leadership backs the AI programme. Now they want to see the value. What do you show them?

You saved 10 hours with AI off the close, or the forecast, or the board pack. Now what?

Hours saved is the laziest way to count

Pankaj Prasoon is Sr Director of Finance Tech & Strategy at LinkedIn. He wrote an article in the Forbes Finance Council with the title: "Hours saved is the laziest way to measure AI in finance."

Counting hours saved as value is the same as booking a receivable as cash. The benefit exists, but if nobody has collected it, it means nothing.

So your 10 hours are a receivable. They are not cash yet.

As a super quick example - In 1972, Southwest Airlines owned four planes and had just lost $1.6 million. So it sold one, for about $500,000. And it kept the whole schedule running between Dallas, Houston and San Antonio, with just the three planes left.

How? Bill Franklin, Southwest's first VP of ground operations, worked out that if each plane stayed at the gate for only 10 minutes, three planes could do the work of four.

Southwest did not get faster and then wonder what to do with the time. It decided first where the time would go (one plane it no longer had to pay for). Then it built the 10 minute methodology.

Time saved only turns into money when somebody decides what it replaces.

So how do you find your ‘4th Plane’ and turn time saved into real results?

Find your 4th plane in 3 steps

1. Time the turnaround: write the before and after for one process

Open the AI ROI worksheet here - ​​Find your AI ROI Worksheet.xlsx​​​

Pick ONE process you do every month. The variance review. The AP run. The board pack.

Write the hours it took before AI in "Hours before AI". Write the hours it takes now in "Hours now".

Say 8 hours before, 6 hours now. That's 2 hours back, and the sheet does the maths for you.

Do it for a handful of processes and you get a number you can present. It is the typical finance method, no?

Tip - Before you count a report, check that someone still needs it. Never automate something nobody needs. Stop sending it on time and see who complains. If nobody complains on the day of the deadline, they didn't need it that day.

Done when: you have one row per process, with the hours back.

2. Give every saved hour a job

Take your hours back from step 1. In column E, give each one one of the following words.

• Never done. Work you always wanted and never had time for. A variance review on every cost centre, not only the top five. A rolling forecast instead of a yearly one.
• Halfway. Work you did, but only as far as the clock allowed. Reconciling the big items and letting the small ones ride. Checking 20 invoices instead of all of them.
• Late. Work that landed after it mattered. The month-end pack on day 8 instead of day 5. The board pack finished on Sunday.

Then, in column F, write the goal. Not "more analysis". Write "variance review on all 40 cost centres". If you can't name the job, the hour isn't saved yet.

An hour with no word next to it is an hour where someone else decides how it's used. Each time you finish early, someone finds you (or your team) a job to fill it.

Done when: every row has one word and one named goal.

3. Fly or sell

Now each hour gets a decision, in column G.

Fly: the hours go to the job you named in step 2.

Sell: the hours replace something you would have paid for. Someone retires and you don't hire a replacement. Write what you no longer pay for, and what it cost.

Sell is the one that needs courage. If AI saves you hours but you still have the same team, you don't save money. You still need the same number of people at closing.

So look at your busiest week, not your average month.

I know Sell is scary, because twice I made my own job disappear.

The first time, I was head of finance transformation. In the reorg, I removed my own role. Some of my tasks moved to other people, and we stopped the rest.

The second time, I was leading a finance team, and our scope got smaller. I gave the leadership to someone in the team, and we didn't need to recruit anyone.

And a smaller team doesn't mean people leave the company. I've seen people move from finance to HR, to procurement, or take a secondment. Finance people know the company and its systems better than almost anyone.

Once you've decided on Fly or Sell, build it into the process.

Basically, there are two ways to use AI in a process.

1. On top: AI helps one person do their own task faster. Like Excel. Most Excel files sit on top of a process. They are rarely part of it.
2. Inside: AI is a step of the process. Take receivables. Instead of a person drafting each reminder, AI drafts it from the client's context and an automation sends it. Nobody has to remember.

So write the AI step into the checklist, with a name next to it in "Owner". Then mark "AI sits" as On top or Inside.

Then brief your team. When work finishes early, they go to the Fly list first.

Tip - Don't move everything inside at once. I tell teams to spend the first 3 to 6 months on top, so people get comfortable. Then move one process inside. And don't expect savings on day one. You grow what your team can do first. The savings come later.

Done when: every row has Fly or Sell, and a name. The boxes at the top of the sheet are what you show leadership.

Bonus Tip - Want it on one page? Ask AI to turn the sheet into an HTML page like the one here.

3 people who found their 4th plane

Mark CFO for 32 years, runs an outsourced CFO practice

We mentioned him already right at the top of this newsletter.

• Before: about 60 hours a week
• After: 15 to 20 hours a week
• Where the hours went: he sold five AI engagements for 2026 (that he’d never done before)
• Fly. A Never done hour (AI engagements), turned into a new line of work.

Caitlin during one of our AI Finance Club Masterclasses

• The job: a financing proposal for a bank
• With AI: it did the research and drafted the entire proposal. It got her to about 80%, and her team finished the rest.
• Saved: "At least 70 hours, probably."
• Where the hours went: the proposal went to the bank, and they got quite a bit of financing from it.
• Fly. As I told her in the course, it was not just a cool market research. It had a real financial impact.

Me, and a letter from the German tax office

• Before: 800 transactions in 80 CSV files that didn't fit the accounting system. One week of work, or about $5,000 to my advisor.
• After: one hour. AI gave me a small piece of Python to merge and clean the files. I have never learned Python.
• Where the hours went: the $5,000 stayed in my pocket. We passed the audit.
• Sell. And the code is now a step in my reporting and my close. It sits inside the process.

See the pattern? None of them stopped at "hours saved". Each one can tell you where the hours went, in one sentence.

The One Thing to Remember

Time that you haven’t protected for high value work will be consumed by whatever shallow work somebody sets for themselves (or is set by somebody else).

So, before you set ‘hours saved’ as the target. Decide what your ultimate goals are, and build towards them instead.

Cash in the bank is always better than an invoice line ;)

And don't forget, if you want me to answer your AI in Finance ROI questions, hit reply, and I will include them in upcoming newsletters.
​

Best,

Your AI Finance Expert,

- Nicolas

P.S. - The best way to learn from me directly is to join my weekly masterclass. 60 minutes + Q&A, where I'll answer all your AI in Finance questions. Join me here.

P.P.S. - Mark found his 4th plane with an AI clone. Here's how to build yours → How to Build Your AI CFO Clone in Claude (2026)​

​