Every ROI number gets litigated. This one is built to survive it.

A net dollar figure and a return multiple, built bottom-up from real sessions, net of the AI Tax.

Your AI ROI · this quarter
Sample data
$1.4M

After AI spend, and after the impact lost to rework.

3.2×
Return multiple
$310K
Recoverable spend
$430K
AI spend
What changes

From an estimate, to a number built out of the org’s own sessions.

Without Tetriz
  • A vendor ROI calculator with your headcount typed into it
  • Seat spend on one side, a feeling of progress on the other
  • A number the CFO can question and nobody can defend
✓ With Tetriz
  • Net effective gain in dollars, and a return multiple
  • Value created and AI spend both attributed down to the team
  • Every line traceable to the sessions and merged PRs behind it
The calculation

How the number is built.

Four steps, in order, each one you can defend.

Cycle-time savings

per merged PR

Loaded engineering cost

your actual cost

AI Tax

rework + review drag

Total AI spend

licence + sessions

Operational AI ROI

net gain + multiple

What it shows

Return, and the places it leaks away.

Net effective gain
Net effective gain · quarter
$1.4M
Value created, net of AI spend and rework
Return multiple
3.2×
Value vs AI spend
Value created
$1.83M
At loaded cost
One team to the whole org

From a single team to the company case.

A team's return, attributed and ranked, then rolled up to the org.

Contributor
Payments team
9 engineers · this quarter
Net gain
$210K
Multiple
3.4×
Recoverable
$38K
Team return
$210K
net gain
3.4×
multiple
$62KAI spend
Value by repo
payments-api
$96K
checkout-web
$71K
ledger
$43K
Returns ranked
1
Payments
$210K net gain
3.4×
2
Checkout
$168K net gain
2.9×
3
Platform
$121K net gain
2.2×
FAQ

Questions leaders ask.

Cycle-time savings per merged PR, measured against the same team's non-AI baseline and valued at your loaded engineering cost, minus total AI spend for the period. Impact lost to rework and review drag is deducted before the net figure, which is why it reads lower than a gross savings number.

No, and the page says so. This is operational AI ROI: the return on the coding workflow, the part Tetriz can build from sessions and shipped code. It's a subset of the engineering org's total AI return, not the whole of it.

Vendor calculators start from seats and assumptions you supply. This one is built bottom-up from your own sessions and merged PRs, so each line traces back to the work it came from rather than an input you typed in.

It's the gap between what AI costs and what it actually ships: review burden, rework, and code that never lands clean. It comes off before the net figure, which is why this number indexes on net rather than gross.

The share of AI spend that went into sessions which didn't convert into shipped output. It's the headroom in the number, and it's usually a harness problem you can fix.