Skip to content
THE RENDER OFFICEAI IMAGE DEPT.

Project budgets

AI image project costs: a campaign budgeting worksheet

The useful number in an AI image budget is the cost of reaching an approved decision. A subscription price or a cost per generated image leaves out rejected candidates, review, and the design work needed afterward. Build the budget around the deliverables you need, then decide how much exploration they justify.

An editorial budgeting worksheet from THE RENDER OFFICE. The amounts and acceptance scenarios below are illustrative assumptions, not customer results or quoted API prices. Current membership and studio controls were checked September 12, 2026.

Count decisions and deliverables separately

Imagine a homeware brand choosing a visual direction for a new collection. The assignment needs one approved campaign concept and three layout drafts. Those are different outputs: a selected image may support several drafts, while several generated images may contribute to one decision. Define what the client is buying before estimating attempts.

List the final work outside the concept session too. Accurate product photography, typography, retouching, mobile crops, and review can remain necessary after a direction is approved. Put an owner and an estimate beside each task so that the generation budget does not silently become the whole production budget.

Build a complete cost ledger

Use the actual membership bill and your own API usage records. THE RENDER OFFICE membership does not include OpenAI API usage. If you subscribe only for this assignment, include the whole monthly fee. If several projects share an existing membership, document the allocation method and keep the full bill visible in the monthly ledger.

For internal planning, time still has a cost even when nobody issues a separate invoice. Choose a consistent hourly cost for the comparison, record hours by activity, and avoid counting the same subscription in both the project estimate and a general overhead allocation.

A project estimate to fill before generating
Cost lineEstimate fromReconcile after the project
MembershipFull bill or documented project allocationActual subscription charge
API usageAn explicit spending assumptionActual usage records for the session
Brief and reviewPeople, hours, and a consistent hourly costTime log including rejected directions
FinishingLayout, corrections, and required exportsApproved production files and actual hours
ContingencyA stated reserve for uncertain workRecord only the reserve actually used as a cost

Work through one transparent example

For illustration, suppose the assignment bears a 490 USD membership fee, 20 USD of API usage, and six hours of work at an assumed 50 USD per hour. The project cost is 810 USD. The API amount and hourly cost are invented budgeting inputs; replace both with your own estimates and reconcile them against actual records.

If that work yields three approved concepts, the cost per approved concept is 270 USD. If only one is approved, it is 810 USD. The total spend has not changed. If none is approved, record no accepted result and the full cost incurred; do not divide by zero or hide the failed exploration in a successful project's figures.

A concept is also different from a final campaign asset. If additional photography or design follows approval, include that work when comparing final deliverables. This calculation does not estimate sales or show that the membership will pay for itself.

Set checkpoints and an attempt ceiling

For the homeware example, the team could reserve an initial group of attempts to compare light and composition, then stop for review. Further attempts should answer a named unresolved question. A count is a planning boundary, not a forecast that the image will become usable by a particular attempt.

Write a stopping rule before the session: stop if the agreed spending or time ceiling is reached, if the concept cannot preserve a critical product feature, or if the team cannot agree on the brief. Moving the remaining budget to photography or design can be a better decision than continuing the same prompt variations.

Close the estimate with actual figures

At handoff, compare estimated and actual membership allocation, API spend, hours, attempts, and accepted concepts. Write one sentence explaining the biggest variance. The next estimate becomes more useful when it records that the brief changed or the client needed extra review, rather than simply recording that AI was expensive or cheap.

Keep this ledger alongside the approved brief. After several comparable assignments you can build your own range of likely costs. Until those records exist, label the range as an assumption and make the first paid project small enough to evaluate deliberately.

Check the full cost against your next brief.

Review the current membership prices and requirements before choosing a plan. Membership renews monthly until canceled, and OpenAI API usage is charged separately.

Review membership pricing