FTE Debt
Baseline the labour already being spent fixing, checking and working around what exists today.
For CFOs and finance leaders
The invoice is visible. The decision risk usually isn't.
You can test a price. You can challenge a return assumption. The harder part is seeing the pattern hidden by the walls of your own company.
What will this replace? What will it cost to run? What has nobody priced because it sits outside the invoice?
The Baseline Package
Before you approve the next AI purchase, do you know your baseline?
Most companies can't show what their last AI purchase returned. Usually that isn't because the tool failed. It's because nobody wrote down the before number, so there's nothing to compare against.
Total Economic Impact, NPV and productivity models all need a real before number: cost displaced, output changed or capacity released. Put in a guess and the formula returns a more precise guess.
Run these three checks as one sequence before the purchase order goes through. Together, they give you the starting point the return case needs.
True Product Cost
Baseline what one data product actually costs to run, including the people and ongoing work around it.
Team Capacity
Baseline what the data team can really take on before another commitment enters the portfolio.
Complete the sequence once, in about a week. Keep the three readings beside the investment case. That is The Baseline Package.
One engagement, measured
€930,000 a year was sitting outside the data budget.
A 120-person company could see its data team and technology spend. It could not see the business hours, unused stack and preventable errors attached to the same work.
€611,000
was stoppable: business hours, unused tooling and preventable errors.
€319,000
was redirectable capacity aimed at work with no decision attached.
3 weeks
to measure it, using the client's payroll rates and cloud bills.
One anonymized engagement. The arithmetic and full breakdown are published on this site.
See the full cost breakdownA purchase order that looked complete
$50,000 in platform credits. Three unanswered questions.
The unit price was clear. The use case was plausible. Nobody had defined the return, the long-term run cost, or whether the controls and underlying data could support it.
Start with the request
Which decision is on your desk?
A new AI or data purchase needs approval
Check what the price and return model cannot tell you about ownership, controls and the data underneath it. Leave with a risk score and the gaps to resolve before approval.
You don't know what data work costs, or whether the team has room for more
Put workaround cost, true product cost and available capacity beside the request before you approve it.
The request does not fit one clean category
Start with the situation on your desk and identify the one existing check built for it.
Keep the map
Get the CFO decision guide.
Four common requests. The right first check for each. One email you can keep for the next budget conversation.