Bonjour ;)

Let me ask you this. How many times this week did you ask ChatGPT something, get back a response, and then straight away think:

"This is way too generic. This is not relevant to me or my company at all.”?

Or how about this?

You spend 10 minutes writing the perfect prompt explaining your role, your company's business model, your current project context. And ChatGPT gives you something that sounds like it was written for any finance person at any company.

So you try again. You add more context, and you explain who your stakeholders are. You describe your reporting structure, and you clarify what "revenue recognition" means in YOUR business specifically.

And 20 minutes later, you've finally got something that works.

But here's the problem. Tomorrow, when you need AI to help with something else, You are starting over from scratch once more.

Over time, all those 20 minutes start to add up to something I call ‘Context Tax’ where you spend more time giving context than you do using AI to improve your work!

Every single conversation begins at zero. With no memory of your company, and no understanding of your voice. With no context about what you're actually trying to achieve.

But what if you just used that 20 minutes to give context once?

And then never had to do this again.

This is exactly what I’m going to show you today - where you will build a ‘Digital Twin’.

​

​
Stop the Context Tax

So let’s say that you need ChatGPT to help you draft variance commentary for your monthly board pack.

This is simple, no?

Except first, you need to explain:

• What your company does
• What level of detail you expect
• How your revenue model works
• What voice and tone is right for your team
• What specific metrics matter to your board
• What happened last month that they'll remember

This is a 300-word prompt. Just to get started.

And even after this, the output comes back sounding like it was written by a generic corporate AI that's never met your team.

Quick maths:

Each time, you are spending:

• 5 minutes writing context
• 10 minutes editing generic outputs to make them specific
• 5 minutes fixing tone/voice issues

That's 20 minutes per task. Times 5 tasks. That's 100 minutes per week.

But here's what makes this really painful.

The more specific your work is, the higher your Context Tax becomes (you have to explain more).

Plus, Context Tax doesn't just cost you time. It costs you consistency.

Because every time you re-explain your context, you explain it slightly differently. You remember different details, and you emphasize different aspects.

So the AI gives you inconsistent outputs.

You're not just paying in time. You're paying in quality.

And, this is really bad because it makes AI feel less useful than it actually is. You are spending so much time on setup and clean up that the actual value gets lost.

​

​
Your Digital Twin

So here's what changed for me.

I stopped trying to give context every time. And I built an AI version of myself instead..

A Digital Twin that knows my company, speaks in my voice, and has my context built into it for all tasks.

And when I need to draft something now? It already knows what it needs.

What is a Digital Twin?

An AI Digital Twin is a personalized version of an AI model (like ChatGPT, Gemini, or Copilot) that has been trained on your specific data, expertise, and communication style.

Note - This is different from the 'Memory' functions in your AI tools. As it is YOU who is specifying the context, not the AI guessing what to remember. I have memory turned off in all of my tools, as this gives me more control, and allows me to be more specific over my AI responses.
On top, you can make several digital twins (one pro and one private, for example).

When you ask a Digital Twin to draft variance commentary, it already knows:

• How you structure your board packs
• What level of detail your team expects
• Which metrics matter most to your exec team
• What tone is appropriate (professional but direct, not corporate fluff)

So instead of spending 20 minutes on a task, you spend 2 minutes. The rest of that time is using AI for what it’s meant to be. An intelligent assistant.

And here is what’s great.

All you need is:

• Your LinkedIn profile (already exists)
• Your CV (already exists)
• A few emails you've written (already in your sent folder)
• Your company website and key documents (already exist)

That's it. Just you, an AI tool you already have access to, and 20 minutes.

I'm going to walk you through the exact process I used. The same one I've taught to hundreds of finance pros in the AI Finance Club.

Some of them are using their Digital Twin to draft variance commentary. Others use it to onboard new team members. One CFO uses it to answer investor questions during fundraising.

But they all have one thing in common: they're no longer paying The Context Tax.
​

​
How to Build Your Digital Twin: The 6-Step Guide

This method uses your existing professional assets to build a knowledge base, plus a set of instructions for a custom AI Digital Twin.

Step 1: Gather Your Professional Data

First, you need to collect the raw materials that define your professional identity.

• LinkedIn Profile: Go to your LinkedIn profile, click "More" (or "Resource"), and "Save to PDF."
• CV/Resume: Have your latest CV available as a separate file.
• Company Information: Use your AI tool (ChatGPT, Gemini, etc.) to scrape your company's website. Use a prompt like:

• Save the output as a document.

Step 2: Define Your Voice

To make sure the AI sounds like you, you must teach it your style.

• Find 3-5 emails you've written that are good examples of your typical writing style (e.g., one to a colleague, one to a boss, one to a client).
• Feed these into your AI with a prompt like:

• Refine the AI's description until it's accurate, and save the final style guide as a document.

Step 3: Consolidate and Upload

Create a new chat and upload all the documents you've gathered:

1. Your LinkedIn PDF
2. Your CV
3. Your Company Info document
4. Your Writing Style Guide document

Step 4: Refine the AI's Knowledge

Now that the AI has your data, you need to fill in the gaps.

Use a prompt like:

Answer these questions in the chat to provide the AI with additional, unwritten context about your goals, priorities, and expertise.

Step 5: Generate the "Master Prompt"

This is the most crucial step. You will ask the AI to create the very instructions you will use to build your digital twin.

Use this prompt:

Prompt (copy/paste)

Step 6: Build Your Custom GPT (or Gem/Agent)

You are now ready to build your tool.

1. Open the builder:

• In ChatGPT: Go to "Explore" -> "Create a GPT."

• In Gemini: Go to "Gems" -> "New Gem."

• In Copilot: Go to "Agents" -> "Create Agent."

2. Configure:

• Instructions: Copy the "Master Prompt" you generated in Step 5 and paste it into the "Instructions" (or "System Prompt") field.
• Knowledge: Upload your 4 source documents (LinkedIn, CV, Company Info, Style Guide) into the "Knowledge" or "Files" section.
• Name & Save: Give your digital twin a name (e.g., "My Digital Twin Advisor") and save it as "Private" or "Only me."

​

​

The Bottom Line

AI should not be painful to use. AI should not become a repetitive task. Building a Digital Twin is the fix. And the beauty of this is, you only have to build it once. Then every conversation after that starts with full context. But this only works if you actually do it. Most finance pros will read this, think "that's a good idea," and then go back to writing 300-word prompts tomorrow morning because that's what they're used to. So here's my question: will you do this right now? ​ Your Move ​ Take 20 minutes right now, follow the six steps, and build your Digital Twin. Then test it, and ask it something you'd normally spend 20 minutes prompting. See if it already knows the context, and see if the output sounds like you. Because here's what I know from training hundreds of finance pros on this. The ones who built their Digital Twin aren't spending Friday afternoons writing the same context over and over. They're actually using AI because it finally works the way it should have from the beginning. ​ Best, Your AI Finance Expert, Nicolas ​ P.S. - What did you think of this approach? Hit reply and let me know if you're planning to try this for your team (I read all replies). ​ P.P.S. - I've got more AI use cases to show in my channel. In this video, I've shown my full AI blueprint on how I train finance teams on AI in 30 days. ​ ​Watch it now!​