5 AI dashboards every finance pro needs by Friday (copy/paste prompts)

Your AI dashboard will fail without 2 things:

Accurate data to de-risk decisions. And a good deck to win over your peers.

Most of you are using Microsoft Copilot, but only 20% of its capability. So on Monday, I will show you how to Master Microsoft Copilot in Excel & PowerPoint. In 60-Mins, you will learn how to:

Clean messy data: Get raw finance data analysis-ready
Analyze finance data: Spot trends, variances, and outliers
Create board-approved slide decks: Turn analysis into a branded presentation

2,604 finance pros registered for my last masterclass with Ramp. As there are not many opportunities to learn from me for free for 60-mins.

You mastered Microsoft Excel already. Now join me here to master Microsoft Copilot.

Master Microsoft Copilot Here

Every week I have around 7 one-on-one coaching calls with finance pros. And most of them are under the same pressure right now: do more, with less.

They've been given initiatives like reduce headcount. And on top, when they ask for budget for a new AI tool, they often get denied. So the question I get, again and again, is: how do I get better results out of the tools I already have?

I give them 2 things.

Number one, use a better model. E.g 'Think Deeper' mode in Copilot will really improve your results.

Number two, use my CSI prompting technique. Context, Specific, Instruction.

So today, let me show you both at once. Here are 5 AI dashboards every finance pro should have live by Friday. I've tagged [C], [S], and [I] inside each one

Let’s go.

1. The board KPI scorecard

This is the dashboard you prepare for your board meeting, built from an Excel file that isn't even formatted.

The one question it needs to answer: are we on track for the quarter?

[C] I'm the CFO of a [type] company, presenting to our board. [S] I need a dashboard that answers one question only: are we on track for the quarter, looking at revenue, gross margin, EBITDA, and cash. [I] Build me an HTML dashboard using the attached trial balance against budget, plus the detail behind it: revenue by product, margin by product, the P&L so you can build an EBITDA waterfall, and the major cash movements. Put the big KPIs at the top, and turn anything off-track red. One screen, no scrolling.

A board dashboard that needs scrolling is not a board dashboard, because people don't have time and people often never go under the fold (past what they see straight away on screen). So keep it simple. Everything on one page.

And of course, always verify your numbers. Ask AI to audit your dashboard against your Excel file before you screen-share this live in a board meeting.

2. The cash runway tracker

This one is for your own team, watching one number every single week: how much runway do we actually have?

[C] I'm a CFO tracking cash runway from our bank export. [S] I need the monthly net burn split into two views: one with exceptional one-time amounts, one without, because a one-off deposit like an office payment shouldn't distort my real burn rate. [I] Calculate the burn, build both views, and if I mention a financing event coming, like funds landing in [X months], factor that in. Set a threshold at [X months] of runway and flag it red if we're getting close.

The trick here is that I'm not just asking for a chart of past data. I'm asking AI to calculate something that doesn't exist in the raw export yet, a projection.

But you have to give it the logic yourself. Explain your own logic, and it will show you exactly when you'd hit your threshold.

3. The budget variance dashboard

This one flags overspend on its own, so you're not trying to look for it every month.

[C] I'm reviewing departmental spend from a file with multiple tabs: cost by department, budget, and actuals. [S] I want one row per department showing budget, actual, variance, and variance in percentage, sorted from worst to best. [I] Build this as a dashboard, flag anything over [X%] or [X amount] with a red, amber, or green signal, and add a filter button for department and cost type.

The trick to name here is simple. Give it the exact sort order and the exact flag threshold in your instruction. Don't leave "overspend" for it to guess.

Bonus - Cost by Department Dashboard

Tomorrow, I'll build a cost-by-department dashboard in Copilot.

First with a "quick/auto" model (bad output), then redone with a "Think Deeper" reasoning model (good waterfall dashboard with filtering).

To see this plus 6 other AI in Finance use cases to help you become an AI CFO this year, join me here.

4. The AR aging dashboard

Use case number four is about money that is owed to you, not money you've spent.

[C] I'm chasing collections from an open invoices export. [S] The real question isn't how old each invoice is, it's who I need to call first. [I] Bucket the invoices into age groups, then rank the clients I should contact first based on amount owed and days overdue.

Tip - Your specific has to name the decision. Say "I want to know who to call first," not "sort by days overdue."

5. The scenario and sensitivity dashboard

This is the most advanced one, and my favorite. A live tool where you drag a slider and watch your forecast change in real time.

[C] Here's my business model and current assumptions: [describe your model]. [S] I want to see what happens to the forecast when price, growth rate, or churn move, live. [I] Build sliders for those three levers. Every time one moves, recalculate and redraw everything, and show my original base case as a dotted line next to the live full line so I can always see what's changed.

I added that dotted line on purpose, because without it, people lose track of what they've changed once the conversation gets going.

And did you notice what I'm doing here? I'm not asking for a static view of data that already exists. I'm asking AI to build a small calculation engine with adjustable inputs. This is a very different prompt than "make me a chart."

The One Thing To Remember

As models become more intelligent, it’s more important that you become specific with your instructions.

Now, the question I always get after something like this: "Nicolas, my data is a mess, this won't work on my files."

Good news. The board scorecard prompt above was built on a trial balance that wasn't even formatted. So you can get going straight away without spending hours on data cleaning.

But on top, always review what comes out before it goes anywhere near a real meeting. AI prepares the work. You're still the one who checks it and owns it.

None of these five dashboards needed you to code. You didn't need a BI team. You didn't even need a bigger AI budget, just a better model and a better prompt.

Best,

Your AI Finance Expert

Nicolas

P.S - The best way to learn from me directly is to join my weekly masterclasses. 60-mins + Q&A where I’ll get all your AI in Finance questions answered. Join me here.

P.P.S - Want to see a live financial dashboard built from scratch? Here's the full build → How to Build INSANE Live Financial Dashboards with Claude

Tips & Insights on Finance & AI