Wednesday, November 11, 12:00 ET. 45 minutes on Zoom. Free.
You'll see:
The dashboard I built with AI, live.
What drives AI usage, and how I keep my work inside my subscription.
The controls a finance team could put in place before the bill takes off.
Hey {{first_name|there}},
This is email 6 of 6 in a series. New here? Start from the beginning:
1. Everyone's talking about AI. Few are using it.
2. My AI disagreed with my EBITDA. It wasn't wrong.
3. The CFO skill nobody admits they're bad at
4. Revenue grew 10%. So why did contribution margin fall?
5. $12M of new sales. $10M of cash gone.
In December 2025, Uber gave AI coding tools to about 5,000 engineers.
To push adoption, it ranked teams by how much AI they used.
By April, the year's AI coding budget was gone. In just FOUR months!
Some engineers were spending $500 to $2,000 a month each.
Uber then capped usage at $1,500 per person, per tool.
Its COO admitted he couldn't tie the spend to the features shipped.
Read that again as a CFO.
Uber rewarded usage and never measured what it produced. The controls came after the money was spent.
Your job is to judge every dollar the business spends and what it brings back. AI deserves the same test. It's just harder to see coming.
Why the bill surprises you
Most AI starts as a subscription. A few seats, a known monthly fee, booked under software. Easy to forecast.
The surprise comes from usage. Once your team connects AI to files, systems and automated workflows, many tools bill by the token: a small chunk of text.
You pay for what goes in: the message, every file, and the whole conversation, which the AI re-reads each time it answers.
You pay for what comes out: the answer, plus the extra "thinking" on harder tasks.
And the most powerful models cost more per token.
Two people can get the same result and run up very different bills.
You don't need Uber's scale for this to hurt a smaller company.
Picture a finance team, fictional but realistic, that automates its month-end review. Each run reads the trial balance, the invoice files and last year's commentary. Someone sets it to run every day.
The bill goes from $300 a month to $4,000.
Nobody approved it. Nobody forecast it. It shows up with the invoice.
And the person who owns cost control in your company is you.
The questions to ask before the invoice does
Who is using AI, and how much?
Who owns AI spend, and which line is it on?
What does each output cost: one month-end review, one board deck?
Is it in next year's forecast?
What's the return, in a number your board will accept?
Your CEO wants more AI. Your board wants proof it pays. Very few finance leaders can answer both.
What I spent
In about 52 hours, on my own time, I built the dashboard, the executive presentation and the financial model you've read about in this series.
I didn't pay a dollar beyond my subscription.
Honestly, I didn't have a choice. My plan had a limit, and I couldn't buy more tokens.
So I had to learn which habits burn tokens and which ones don't.
That limit is the control Uber added after the money was gone. I had it from day one.
See it live on Wednesday
On November 11 at noon ET, I'll show the dashboard I built with AI, live, and the controls I'd put on AI spend before the bill takes off. 45 minutes on Zoom, free.
P.S. Do you know what your company spent on AI last month? On Wednesday I'll show how I keep mine inside a fixed subscription. Save your seat here.
Disclaimer:
This content is for informational and educational purposes only and should not be construed as financial, legal, or professional advice. Always consult with a qualified advisor before making any business or financial decisions. The author and publisher disclaim any liability for actions taken based on this content.
Talk soon,
