You waste 3 days gathering data and consolidating 37 Excel workbooks. And still end up in super awkward meetings where the data is out of date already – You lose motivation, and your brain ends up fried.
But, it’s 2026, and you should not have to deal with this.
So, to help solve this I’m running this free 60-minute masterclass for you with Spendesk. You will see:
- The one prompt that surfaces every aging bucket, supplier exposure, and stuck invoice over €10K (in seconds)
- How to turn a month of messy expense data into clean, categorized entries.
- How to build a board-ready dashboard from raw data (without the Sunday evening Excel session)
So, if you want to stop wasting time answering questions on your own numbers. And start spending more time on the things that give you energy and make an impact – reserve your seat here before we’re full.
| Reserve My Free Masterclass Seat |
The finance lead of a $70M business just turned his team of 4 into a team of 5
Last week I had a call with Marcus, one of our AI Finance Club members. He's the in-house finance lead at a services business doing around $70M in revenue, running a team of four.
He was almost laughing on the call. It was like a "I can't believe this worked" laugh.
He'd spent a weekend building his AI Clone in Microsoft Copilot using the 7-step method I teach in the masterclass.
He thought he'd get a glorified email drafter. But, what he got was a fifth team member!
The payroll accrual file that used to make him want to cry every month? His clone now handles the first review. The recurring "how do you think about X" questions from his team of four? They're drafted in a much less time.
Plus, he's now running weekly AI lunch-and-learns to bring his team along – because, as he said to me, "teaching is learning twice."
Today I'll show you the same 7-step playbook Marcus used. For you to build your own AI clone in 90 minutes.
Coinbase just flipped its org chart cutting 14% of its team.
|
|
Roughly 700 Coinbase employees recently lost their jobs, and Brian Armstrong said this is because he’s flipping his org chart upside down. He's rebuilding the company around ‘AI-native pods’ with one-person teams running agent fleets.
– Armstrong on X – Full update here.
Industry estimates suggest almost half of the 92,000 tech layoffs in Q1 2026 were tied to AI-driven restructuring [Economic Times here].
Coinbase is proof that this trend is real.
This means you need to be asking this question "what does our finance org look like when one analyst can do the work of three?"
It is really important that your answer is better than "we’ve given everyone ChatGPT license," because the teams that will win in the next 12 months will have leaders that have already codified their own thinking into something the rest of the team can use.
Your AI Clone is the smallest, quickest thing you can start doing to build an AI-native finance function.
And it’s your knowledge (and your ability to systematize) that will make you irreplaceable.
Yes, it saves you time. But it also makes other people more knowledgeable and capable as a result.
Now, you need to think of yourself as an orchestrator.
7-Steps to Building Your AI Clone
The process starts with your LinkedIn profile, exported as a PDF.
That one document holds 10+ years of how you describe your work. Your roles, sectors, achievements, and it's written by you (in your voide) for the public.
From there you layer on: company context, three real email samples for tone, a self-interview the AI runs on you. Followed by a generated system prompt under 8,000 characters.
Here’s the detailed breakdown.
1. Hand it your LinkedIn:
Open LinkedIn, click "More," choose "Save to PDF." Download. This is your foundation file.
|
|
2. Add the supporting docs only you have:
|
|
Resume, bio, any keynote brochure or services one-pager. PDF them. Five files max to begin with (quality over volume).
3. Add in Company Context
|
|
Save the output as a PDF.
4. Capture your tone using 3 real emails:
One to your CEO, one to a board member, one to a peer.
Paste them in. If you don't have samples, just enter the prompt:
5. Let the AI interview you back:
Upload everything and prompt:
Answer them, then repeat 2-3 times until it stops surprising you.
6. Generate the system prompt (under 8,000 characters):
Save the output.
7. Deploy where you work:
- ChatGPT → Create a GPT.
- Copilot → New Agent.
- Gemini → New Gem.
- Claude → Project.
Paste the system prompt, upload your PDFs and save.
Pro tip: Tell the clone to ask you when it isn't sure, rather than guessing. This will help reduce the chances of it getting anything wrong.
From 45 minutes to 5 minutes – The Google Form Use Case
|
|
Last Wednesday I needed to build a Google Form to capture feedback from our weekend AI Finance Club team event.
Manually, that's about 45 minutes of work to set up the questions, add multiple-choice options, the logic, the descriptions etc.
Instead I had a voice-chat with my own AI Clone in ChatGPT (which already knows the business, my audience, and the format).
And I asked it to give me the prompt to use with Gemini to create the form.
The form was done in about 5 minutes.
9 times faster on a specific task that I would have otherwise done in the evening when I should be spending time with my daughters.
Plus, don’t forget, all of these compounds. One small thing replaced leads to another, leads to another.
Before you know it, all of these small automation investments will add up, and clear your schedule for the work that gives you energy and creates a big impact!
The One Thing to Remember
Your clone is just the first ‘node’.
Marcus is already at the second node – his weekly AI lunch-and-learns are to bring his team of four along with him – because once you've built one AI Clone, the method scales fast.
Plus, the context documents you create once become reusable across every clone you ever build.
Coinbase calls them pods. Marcus calls them his team. You can call them whatever you want.
What matters is who builds the first node – and right now, in your finance function, that has to be you.
Build it tomorrow. Then run an AI lunch-and-learn with your team next week.
Then reply to me with what you used it for first. I read every reply.
Best,
Your AI Finance Expert,
– Nicolas
P.S. – Unsure where to start with AI? Reply to this with your biggest challenge. I will try and help where I can.
P.P.S. – Built your clone? Here's the 30-day playbook to bring your whole finance team along with you → How to Train Your Finance Team on AI in 30 Days