Monitor AI credit usage

What this achieves

AI credits are consumed every time an AI feature runs, and this screen shows how many you have, how quickly they are going, and where they are going. It tells you whether the allowance will last the month before it runs out rather than after.

Required role: Administrator.

Steps

Go to Administration → AI Governance → AI Credits.
Choose the Period.
Read the four figures, then the breakdowns beneath them.

The four figures

Option	Description
Included	Credits included in your plan.
Remaining	Credits still available.
Total used	Credits consumed in the selected period.
Days until depletion	How long the remaining credits last at the current rate.

Days until depletion is a projection from the recent rate of use, not a countdown. A quiet week extends it and a busy week shortens it, so read it as an early warning rather than a date.

Where credits are close to running out, the screen says so directly, naming the number of days at the current rate.

Example: HC Corp sees 14 days until depletion in the middle of the month. That is the moment to look at the breakdowns and decide whether the consumption is worth it, not the moment to wait and see.

The period only moves some of the figures

The Period selector covers recent months, and it changes what is measured — the usage, the breakdowns, and the request counts.

Included and Remaining describe your plan and your current balance. They are not a historical figure for the month you selected, so do not read the four figures as one consistent snapshot of a past month.

Where the credits went

Option	Description
By module	Consumption by part of the product.
By feature	Consumption by individual AI feature.
Daily trend	Consumption day by day across the period.
Top users	The people consuming most, with their credits and request counts.

By feature is the breakdown that usually answers the question. Module tells you which area; feature tells you exactly which action, and that is the level at which you can decide whether it is worth the spend.

The Daily trend is where a one-off spike is distinguishable from a rising baseline. Those need different responses — a spike is usually somebody trying something out, a rising baseline is adoption.

Top users is a usage figure, not a performance one

The top-users list names people and counts their consumption. Somebody high on that list is using the features they were given, which is what the allocation was for.

Treat it as a cost breakdown. Where consumption looks disproportionate, the question is whether the feature is the right tool for what they are doing, and that is a conversation rather than a conclusion.

Example: Priya Raman appearing at the top of HC Corp’s list means she uses AI most. Whether that is good depends entirely on what she is using it for, which the prompt log will tell you and this screen will not.

Running out

When credits run out, AI features stop working for the people whose allowance is exhausted. The screen points you at your account manager to change your plan.

Two things are worth doing before that point. Check the by-feature breakdown for consumption that is not earning its keep, and check the allocations — credits are given to individual employees, so a large unused allocation sitting with somebody who never uses AI is capacity you already have.

What happens next

Credit monitoring is a budget control that feeds two decisions: whether to change your plan, and whether to change who has an allowance. Where you need to know exactly what a figure was spent on, the prompt log carries the per-call detail behind every credit counted here.