WORKFLOW • Shot-budget worksheet added · Jul 24, 2026
AI Video Storyboard Budget: Estimate Credits by Shots, Not Final Runtime
A 30-second edit may contain six shots or sixty, and the credit requirement can be completely different. Final runtime is a weak budgeting unit because AI video is generated and approved one shot at a time. Start from the storyboard instead: count the unique shots, identify which ones need variants, and apply an honest attempt rate to calculate the generation volume before a project begins.
Turn the storyboard into a shot inventory
List every distinct generated shot in the cut rather than every editing beat. For each one, define the target length, aspect ratio, model route, and whether a reference image or continuity constraint is required. A two-second transition that must match the product and lighting can cost more to solve than a ten-second atmosphere shot. Give each shot an owner, a priority, and a fallback so the budget reflects the work that truly needs generation rather than a vague total runtime.
Separate master shots from format derivatives
Do not automatically multiply all shots by every output format. Some vertical, square, and landscape exports can come from a master composition, while others need a newly framed generation to protect the subject or product. Mark each shot as reuse, crop, or regenerate for each destination. This makes format expansion visible: a campaign with one master film and five platform variants may require a few additional generations or an entirely separate set of vertical shots. Budget the latter as a new shot pack, not a free export.
Apply an attempt rate by shot difficulty
Use a different attempt assumption for simple and difficult shots. A basic texture or scene-establishing shot may be approved quickly; character continuity, product fidelity, readable on-screen text, or precise motion normally deserve a higher retry allowance. Multiply each shot's planned generations by its expected credit cost, then add the groups together. If you only use an average across the project, difficult hero shots can quietly consume the buffer intended for everything else.
Keep a protected revision reserve
Client feedback and internal review create another generation cycle after the first assembly. Reserve a share of the budget for changes to approved work rather than spending every credit in the initial pass. The reserve should be tied to the number of decision-makers, brand sensitivity, and the number of versions promised. If the reserve is unused, it becomes evidence to reduce the next project's plan size; if it is routinely exhausted, update your storyboard assumptions instead of treating overage as an exception.
Build a shot-budget worksheet before generation starts
Use one worksheet row for every planned generated shot, including format derivatives that require a separate generation. Record: shot ID, use in the edit, master or derivative status, target duration, aspect ratio, model route, current provider rate, planned attempts, revision reserve, and fallback. The row formula is simple: planned generation units equal planned attempts plus revision reserve; planned credit cost equals planned generation units multiplied by the provider's current rate for that exact model, duration, and setting. Sum the rows by shot group rather than relying on one project-wide average. Keep a separate column for reuse, crop, or regenerate so a derivative that can reuse a master adds no new generation units, while a reframed vertical shot is visible as a new cost. This worksheet is a planning artifact, not a promise of output quality: update it with actual attempts after production so the next storyboard starts from observed usage.
Reconcile the forecast after the first assembly
When the first edit is assembled, compare every worksheet row with what actually happened. Record actual attempts, approved output, credits used, whether the planned fallback was needed, and the reason for any variance. Classify the difference as a changed brief, a harder-than-expected shot, a format regeneration, an avoidable retry, or unused reserve. Do not hide those differences inside one project total. Row-level reconciliation shows which storyboard assumptions should change next time and which overage came from a new decision after production began. It also gives the review team a factual basis for protecting the remaining reserve before requesting another generation cycle.
Editorial note: This framework is general information, not a vendor endorsement. Check the current pricing, terms, and data-handling details directly with the provider before buying.