6
Every finance team has that task.
The one you put off until you can clear a big block of time, pour a fresh cup of coffee, and mentally prepare.
For Joanna Rossomano, that task is building the monthly budget allocation for the year ahead.
If you have ever done this kind of work, you already know why.
To prepare the 2026 monthly allocation, Joanna typically needs to:
- Start with annual totals by expense code
- Look back at 2025 monthly spending patterns
- Carry those patterns forward into 2026, expense code by expense code
- Respect how each account actually behaves
- Some are straight‑line
- Some are seasonal
- Some spike once a year
- Account for department changes since last year
- Ensure everything ties out perfectly across hundreds of rows
This is not simple distribution.
It is a year‑over‑year behavioral rebuild of the budget.
Manually.
Line by line.
Joanna does this work exceptionally well. It is also the kind of work that quietly consumes time and focus every budget season.
And it is the kind of work that makes you think,
“If only I could ask someone to help me with this.”
Setting the Stage: A Real Test, Not a Demo
When Joanna and I met, the goal was not to talk about AI in theory. It was to test something real, using her actual data and her real workflow.
She already had what she needed:
- Her 2025 monthly allocation baseline
- Her 2026 annual totals by expense code
What we wanted to understand was simple.
Could she describe the work the same way she would to a coworker, and let Agent Mode handle the heavy lifting?
Before she typed anything, I explained one important shift.
Agent Mode is not about thinking harder about prompts.
It is about thinking more naturally about how you ask for help.
So we started with one sentence.
The Sentence That Started It All
Joanna did not write a formula.
She did not explain steps.
She did not tell Excel how to do the work.
She simply asked for help the way she would ask a coworker.
“Using the 2025 Allocation (Baseline) worksheet, create a new worksheet that allocates each 2026 annual total across 12 months using the same monthly distribution pattern by expense code.”
One sentence.
Rows upon rows of data.
Work that would normally take hours.
Then she hit Enter.
Watching It Work Changed Everything
Within seconds, Agent Mode began working through the request. Joanna scrolled quietly as the worksheet filled in.
Then she said it.
“That’s crazy.”
Not because the tool was doing something unfamiliar, but because it was doing the work the way she would have done it herself.
The results spoke for themselves:
- 294 of 306 rows matched by department and account
- Straight‑line accounts stayed straight‑line
- Seasonal accounts kept their curves
- One‑time expenses landed logically
- Missing historical patterns were filled evenly
This was not guesswork.
It reflected her budgeting logic.
Then she noticed the account she expected to cause problems.
Travel.
The Travel Test: Where It Usually Falls Apart
Travel is normally the account that takes the longest.
It is not linear.
It changes year to year.
People move departments.
Joanna assumed this would be where things broke.
Instead, Agent Mode:
- Picked up the 2025 seasonality
- Applied it to the 2026 totals
- Adjusted for department shifts
- Filled all 12 months automatically
She paused.
“This would have taken me forever.”
What mattered even more was what came next.
She realized she now had a complete baseline.
Not something she had to build from scratch, but something she could review, validate, and adjust.
The same approach she’d take if she’d asked someone on her team for help.
When the Visuals Changed How She Thinks
Once the allocations were complete, we asked Agent Mode to create visuals to help create a quick-scan view for her validation process.
Joanna had never done this before, not because she couldn’t, but because it would take too much time.
The charts appeared instantly.
- Monthly distributions by account code
- Clear seasonal patterns
- Spikes and anomalies that stood out immediately
She leaned in.
“This is going to save us so much time.”
Instead of filtering one account at a time and scanning cells, she could see the story of the budget.
That is when the lightbulbs started going off.
From Task to Teammate
Without any prompting, Joanna began thinking ahead.
- Could I compare 2025 actuals to our 2026 plan visually?
- Could it flag unusual spikes or dips?
- Could this help with vendor and contract reviews?
- Could I reallocate hiring fees as assumptions change?
At that moment, something shifted.
She was no longer evaluating a tool.
She was imagining a teammate.
Someone who could take on the heavy lifting so she could focus on analysis, validation, and decision‑making.
That’s when I told her to think about Agent Mode very literally.
Ask it the way you would ask a coworker on your finance team.
In natural language.
Without overthinking it.
That is where it really clicked.
This Is Not Just About Speed
Yes, Joanna saved time.
Yes, the allocations were right.
Yes, the visuals changed how she reviews the budget.
But the bigger impact was this.
She now starts from a complete draft instead of a blank sheet.
She validates instead of rebuilding.
She analyzes instead of assembling.
And she gained back time and mental space during one of the busiest parts of the year.
Most importantly, she gained something she never had before.
A teammate for the hardest part of budget season.
Want to Try What Joanna Did?
The March 2nd Copilot Tip of the Week covered Agent Mode in Excel and walks through the basics so you can try a similar approach with your own data.
👉 You can find the Copilot Tip here.
If there is one takeaway from Joanna’s experience, it is this:
You do not need a clever prompt.
You just need a clear request.
That is where these digital teammates shine.
What could Agent Mode help you with?
Let’s keep learning, experimenting, and sharing.
Doug
