Give the AI the locked script, the locations, the kit list and the crew's hours, and ask for a shot list in script order that adds nothing you cannot shoot. Regroup the shots into set-ups, then draft a timed schedule with buffers and a call sheet. The director checks coverage, light, access and the arithmetic before anyone is booked.
A generator's weak spot is the day, not the list. It will schedule a morning exterior after lunch, ask for a drone shot when you own no drone, and total a schedule that forgets moves and meals. Shot-list pages stop at the list. The director's checks are what keep the day on time, and a 90-second corporate film for a print shop runs through all of them below.
Give the AI four lists, not just the script
A script says what the audience sees. A shoot needs four more things written down: where you can film and when, what kit you own or have hired, who is available and for how long, and the day itself. Without them the assistant fills the gaps with a generic film crew that does not exist. Paste all five into one prompt.
You are a first assistant director helping me plan one shoot day.
Use ONLY the inputs below. Do not add a shot that needs kit, people or a
location I have not listed.
SCRIPT: [scenes with voice-over lines and timings]
LOCATIONS: [area, access time, power, noise, permissions, spare rooms]
KIT: [camera, lenses, tripod, slider, gimbal?, lights, microphones]
PEOPLE: [crew roles and hours; talent and their availability windows]
DAY: [general call, hard wrap, sunrise and sunset I looked up]
TASK 1: Numbered shot list in script order. Columns: Shot | Scene |
Size and angle | Movement | What happens | Location | Kit | Seconds.
The seconds must add up to the script length.
TASK 2: Mark any shot that needs something not listed as NEEDS: [item].
TASK 3: Group shots into set-ups (same position and lens) and list the
shots in each set-up. Do not schedule yet.Look up sunrise, sunset and the sun's direction on the shoot date yourself. The assistant does not know where your shopfront faces, and a confident wrong answer about light is worse than none. If the script is not locked, finish it first; my tutorials on an AI video script generator from a brief and an AI storyboard generator from a script cover the steps before this one.
The 90-second film we will plan
The client is a print shop, and the film shows how an order is made. The shop, the script and every figure in this worked example are illustrative. The kit list is one camera, a 35mm and an 85mm prime, sticks, a slider, two LED panels, a lavalier and a shotgun microphone. There is no gimbal and no drone. The crew is three: director, camera operator, sound and assistant. The script has seven scenes (illustrative):
| Scene | Time | Content | Location |
|---|---|---|---|
| 1 | 0:00 to 0:08 | Owner opens the shop for the day | Exterior |
| 2 | 0:08 to 0:22 | Customer and account manager agree an order over a printed proof | Front counter |
| 3 | 0:22 to 0:38 | Designer checks the proof on screen against colour swatches | Prepress office |
| 4 | 0:38 to 0:58 | Press running, operator inspects a sheet | Press floor |
| 5 | 0:58 to 1:10 | Cutting, folding and packing | Finishing area |
| 6 | 1:10 to 1:22 | Owner to camera: "We check every sheet before it leaves." | Meeting room |
| 7 | 1:22 to 1:30 | Customer collects the order; end card | Front counter |
The Complete Excel Course with AI Integration
From formulas and pivot tables to Power Query, DAX, VBA and AI with Copilot, ChatGPT and Claude — hands-on from the first lesson.
By Sagnik Bhattacharya and Virginia Thorn, Primetime Emmy Award winner
Buy nowThe first draft and the fixes that mattered
The first draft came back with 34 shots totalling 96 seconds across 12 set-ups, and a day of about seven and a half hours that looked reasonable. Here is the top of the cleaned list, eight of its 32 shots, after the fixes below.
| Shot | Scene | Set-up | Size and angle | What happens | Kit | Secs |
|---|---|---|---|---|---|---|
| 1A | 1 | S1 | Wide, locked off | Shop front at opening; owner walks to the door | 35mm, sticks | 4 |
| 1B | 1 | S2 | Close-up, locked off | Hand turns the key; sign flips to Open | 85mm, sticks | 2 |
| 2A | 2 | S3 | Medium two-shot | Customer and account manager, proof between them | 35mm, sticks | 4 |
| 2B | 2 | S4 | Close-up | Finger traces a line on the printed proof | 85mm | 3 |
| 3A | 3 | S5 | Medium, over the shoulder | Designer checks the proof on the monitor | 35mm | 4 |
| 3B | 3 | S5 | Close-up, locked off | Colour swatches fanned beside the screen | 85mm | 3 |
| 4A | 4 | S6 | Wide | Press running, operator in frame | 35mm, sticks | 5 |
| 4B | 4 | S7 | Medium | Operator lifts a sheet and checks it under light | 35mm | 3 |
Four problems turned up in the draft, and each shows what to look for in yours.
- Kit it did not have. Shot 0A was an aerial reveal of the shop, and the press-floor walk-through was a gimbal move. The NEEDS tag flagged the aerial. The gimbal one slipped through because the assistant described it as "smooth tracking". The aerial went, and the walk-through became a locked-off wide.
- A shot from the wrong time of day. A night exterior with the shop lights glowing did not belong in a film that opens at morning. It was 3 seconds, and it went.
- Screen time. The seconds column added up to 96 against a 90-second script. Removing the aerial and the night exterior, three seconds each, brought the total to exactly 90 across 32 shots.
- A coverage gap. The voice-over line "We check every sheet before it leaves" had no matching pictures. I swapped two finishing close-ups for cutaways of sheets being inspected, which cost nothing in count or time.
Coverage is the check the assistant cannot do for you. Read the voice-over line by line against the shot list and write the shot number beside each line. Any line with no number is a hole in the edit that you will discover in post-production when it is expensive.
The draft's wording also needed translating into things a crew can shoot. Assistants write mood where a camera operator needs a position, so rewrite each vague line into size, lens, movement and duration.
| The draft said | Shootable version |
|---|---|
| "Dynamic shot of the busy press floor" | "Locked-off wide, 35mm on sticks, press running, operator at the right third, 5 seconds" |
| "Emotional close-up of the owner opening up" | "85mm close-up of the key turning in the lock, 2 seconds" |
| "Smooth tracking move through the shop" | "Two locked-off wides, one from each end of the aisle, cut together in the edit" |
| "Cinematic shot of the proof" | "85mm close-up, window light from the left, a finger traces one line, 3 seconds" |
Group the set-ups, then do the arithmetic the AI skipped
Shots do not cost time; set-ups do. Moving the camera, relighting and checking sound is the expensive part, and a handful of shots from one position is cheap. The 32 shots fell into 12 set-ups, and the second prompt turns those into a day.
Using the approved shot list and set-up groups below, draft a shoot-day
schedule. My assumptions: simple set-up 20 minutes plus 10 minutes of
shooting; interview set-up 35 minutes plus 20 shooting; 20% buffer on
each block; 10 minutes to move between areas; 45-minute lunch;
30 minutes load-in; 30 minutes wrap. General call [time], hard wrap [time].
Order by location, access and light, not script order. Show start and end
for each block and the total length of the day. If the day passes the
hard wrap, name the shots that give least and propose cuts.Those figures are planning assumptions of mine, not industry standards, so replace them with your crew's real averages after a few shoots. Run the sum yourself, because the draft schedule you get back is the one most likely to be wrong.
| Item | Minutes | Running total |
|---|---|---|
| 11 simple set-ups at 30 minutes (20 set-up, 10 shooting) | 330 | 330 |
| 1 interview set-up (35 plus 20) | 55 | 385 |
| Load-in and kit check | 30 | 415 |
| Wrap | 30 | 445 |
| 20% buffer on each block | 77 | 522 |
| 5 moves between areas at 10 minutes | 50 | 572 |
| Lunch | 45 | 617 |
The assistant's draft day was the 445 minutes in the fourth row, about seven and a half hours, because it left out buffers, moves and lunch. The real day is 617 minutes, or 10 hours 17 minutes. With a 07:30 call and a 17:30 hard wrap that is 17 minutes over, so something has to give. Merging the two finishing set-ups into one wider position saves 36 minutes with its buffer, giving a 9 hour 41 minute day and a 17:11 wrap. Set-ups fell from 12 to 11, and nothing the script needs was lost.
Rank the shots so cuts are decisions, not accidents
Add one column to the shot list for priority. Ask the assistant to propose A (the edit fails without it), B (the edit is better with it) and C (nice to have), then overrule it where you know the cut. In the print shop plan 20 of the 32 shots are A, 8 are B and 4 are C. The C shots are two roller close-ups on the press floor, a second angle at the counter and one finishing cutaway. If the press block overruns by up to 36 minutes, which is one set-up plus its buffer, you drop the roller close-up set-up and lose nothing the edit depends on. Without the column, the cut is made at half past two in the afternoon by whoever is most tired, and it often removes the interview. Treat priority as part of the shot list, not a footnote, and carry it onto the call sheet.
Check the schedule against light, access and people
Once the arithmetic works, test the order against the real-world limits. The final schedule, illustrative:
| Time | Block | Why it sits there |
|---|---|---|
| 07:30 to 08:00 | Load-in, kit check, safety briefing | Before the shop opens |
| 08:00 to 09:12 | Exterior, 2 set-ups | Sun reaches the shopfront in the morning (from the lookup); also matches the script's opening |
| 09:22 to 11:10 | Front counter, 3 set-ups (scenes 2 and 7) | The customer is only available 09:30 to 11:30, so both counter scenes go together |
| 11:20 to 11:56 | Prepress, 1 set-up | Quiet; no dependency |
| 12:06 to 12:51 | Lunch and card offload | First backup of the cards |
| 12:51 to 14:39 | Press floor, 3 set-ups | The operator can run the press only until 15:00 |
| 14:49 to 15:25 | Finishing, 1 set-up | Follows the press, as the work does |
| 15:35 to 16:41 | Interview, owner to camera | Meeting room is quiet once the press has stopped |
| 16:41 to 17:11 | Wrap, second backup, kit check | Ten-minute moves sit between each block |
The first draft had scene 7, the customer collecting the order, last thing in the day, when the customer was long gone. That is the kind of constraint an assistant misses unless the availability window is in the prompt, and even then it is worth checking. Another practical fix: the owner appears in scene 1 at 08:00 and in the interview at 15:35, so I would photograph their jacket and collar at the first set-up to keep the continuity.
Ask the AI to attack its own plan
Once the plan balances, use the assistant as a sceptical line producer. Paste the schedule and ask: List the ten most likely ways this day fails. For each give the evidence in the plan, how likely it is (low, medium or high) and the cheapest fix. An illustrative answer for the print shop included three items worth keeping. The press operator's job runs late (medium), and the fix is to swap the press and finishing blocks, which costs no time because both are in the same hall. The owner's two interview lines take far longer than planned (high), so send the lines the day before and keep the 66 minutes already in the plan. Rain on the exterior (medium) is answered by the canopy fallback. The assistant also raised four items I judged unlikely and dropped. Keep the survivors in the call sheet notes so the crew knows plan B before they need it.
The checks that sit outside the schedule
- Sound. The draft mentioned none. Add 60 seconds of room tone at each area, put the interview in the meeting room away from the press, and decide who is on audio at every block, not just the interview.
- Safety near machinery. Agree a short risk assessment and briefing with the shop's floor manager before filming near the press, plan hearing protection if it is loud, and tell the crew where not to stand. Check any drone rules and pilot requirements before a drone is ever in a plan.
- Permissions and releases. Get written permission for the locations and a signed release from every person who appears. Confirm which of the shop's staff the client is happy to see on camera.
- Storage. The cameras will roll for about five minutes per set-up and ten for the interview, so 60 minutes in all. At an example bitrate of 100 megabits per second that is 750 MB a minute, or 45 GB, and 90 GB with a backup copy. Read your camera's real setting, then check that the cards and drives hold it.
- Weather. Name the fallback for the exterior: shoot it under the entrance canopy or move it to the end of the day if there is better light, and tell the client in advance.
- Spare capacity. If the day has no slack, keep two shots flagged as droppable so the last block can still finish when something runs over.
The call sheet, filled in
With the schedule agreed, the assistant can lay it out as a call sheet: paste the final schedule and the contact list and ask it to format without changing any time. Check the times against your own table afterwards, and fill in the bracketed items by hand. For this shoot (illustrative):
- Production: [client name], "How your order is made", 90-second corporate film.
- Date and calls: [shoot date]. General call 07:30, safety briefing 07:45, hard wrap 17:30.
- Location: [address as agreed in writing], parking [details], power [outlet locations], nearest medical facility [name and address].
- Light: sunrise and sunset [looked up for the date and place], sun on the shopfront from [time from the lookup].
- Cast calls: owner 07:45; account manager 09:15; customer 09:30 (window ends 11:30); designer 11:15; press operator 12:45 (press available until 15:00); finishing operative 14:45.
- Crew: director, camera operator, sound and assistant, all 07:30.
- Notes: hearing protection on the press floor; releases signed on arrival; 60 seconds of room tone per area; cards offloaded to two drives at lunch and at wrap; exterior fallback under the canopy if wet.
Send it the day before, and read it through once as if you were the customer who has to arrive at 09:30. A call time wrong by half an hour costs more than the planning did. On the day, use the same shot IDs on the slate, saying "2B, take 3" aloud, so the editor can match every clip to the script without guessing.
The same method for other kinds of shoot
The constraint lists change with the sector, and so do the questions the director has to ask. For a tutoring centre's recruitment film, children on camera mean written consent from parents or guardians before the day, filming around lessons so no class is interrupted, and nothing on screen that identifies a pupil. For a yoga studio's class film the room is the constraint: it is free only between classes, daylight shifts across the slot, and street noise shows up on the track, so the shot list is short and the schedule is a series of 45-minute windows. For a wedding planner's showreel, nothing can be restaged, so the plan leans on a second camera and a priority column built around the moments that happen once. In each case the method is unchanged: four lists, set-ups, arithmetic with buffers, then checks against light, access and people.
Where dedicated tools fit, and what the planning costs in time
A spreadsheet carries a plan like this one without difficulty, and it is easy to share with a client. Dedicated production tools add templates and layouts. Filmustage's pricing page, for example, lists storyboards and shot lists, schedules and call sheets as free, non-AI features, and its free plan includes partial AI breakdown of a script, covering 20% of scenes, in one active project. The paid Director's Cut plan starts from $49 a month, or $408 a year billed annually. The page lists AI features for script breakdown, scheduling and budgeting but no AI shot list, so the shot list stays a human-checked step whichever tool holds it. My tutorial on AI filmmaking tools by job compares more of them.
My working estimate for a shoot like this one is about 75 minutes: 10 to gather the four lists, 20 for the AI to draft the shot list and the schedule, and 45 for the director's review and arithmetic. From scratch I would expect two to three hours, and the checks above are where the time saved is most likely to be lost, so do not skip them to save ten minutes. Once the day's length is known, it feeds directly into what you charge; my tutorial on pricing videography by day rate, project fee or package shows how, and an AI rough cut from a transcript is the step that follows the shoot.
Questions producers ask about AI shot lists
Can AI make a shot list from a script?
Yes, if you supply the script, locations, kit and crew hours and tell it not to add shots that need anything missing from those lists. It drafts in script order. A director then regroups the shots by set-up, checks coverage against every voice-over line and recomputes the schedule, because a first draft usually ignores moves and buffers.
How many shots does a 90-second film need?
There is no fixed number. At an average of about three seconds a shot, 90 seconds needs around 30, and films built on slower interviews need fewer. Plan a few alternates for the edit, then count set-ups, because set-ups rather than shots drive the length of the shoot day.
How long should I allow for each set-up?
Set your own figures from past shoots. A starting point is 20 minutes to set up a simple position and 10 to shoot it, 35 and 20 for an interview, plus a 20% buffer on each block. Time your first few shoots and replace these numbers with your crew's real averages.
Should I use a dedicated shot-list tool or a spreadsheet?
A spreadsheet is enough for most corporate shoots and is easy to share. Dedicated tools add templates and call sheet layouts. Filmustage's free plan, for example, lists shot lists, schedules and call sheets as free features but limits AI script breakdown to part of a script and one active project, so check what you would use.
What does AI usually miss in a shoot-day plan?
Moves between areas, meal breaks, access times, kit it cannot see, talent availability windows, sound and safety near machinery. It also gets light wrong because it does not know where the sun sits at your location on your date. Look up sunrise, sunset and sun direction yourself.
Further reads
- Video Production Proposal: Scope, Budget and Timeline Drafted With AI — Turn the shoot plan into a proposal with scope and budget.
- Video Production Pricing Sheet: Build a Rate Card With AI — Price the shoot day from the schedule you have just built.
- How Can AI Help Videographers Consolidate Client Revision Feedback? — Consolidate client feedback after the edit.
- AI Video Production Software Compared for a Small Studio — Compare the wider AI tools a small studio might add.
- AI Subtitle Generator for Client Videos: Accuracy and Delivery Checks — Check subtitles before delivery.
- Videographer Software: Client Review, Approvals and Invoices Compared — Handle review, approvals and invoices in one tool.
- Best AI for Filmmakers Selling Video Services to Local Businesses — A tool-by-stage guide for video producers with local clients: scripts, shot lists, rough cuts, monthly costs and the client-work checks that creator lists skip.
- Free AI Video Generator for Small Business Ads: Limits to Check First — Ten free-tier limits to check before a small business builds an ad on a free AI video generator, with a credit sum, a script and a shot list.
- AI Tools for Filmmakers: Script Breakdowns, Shot Lists, Call Sheets — Assistants or production platforms for breakdowns, shot lists and call sheets on a small paid shoot, with worked examples and what to check by hand.
- AI Video Content Agency: Services a Small Studio Can Add and Price — A six-service menu for a small video studio, a worked pricing floor, and the rights, consent and disclosure checks before delivery.
- AI Video Translator for Corporate Video: Dubbing, Subtitles, Checks — Choose subtitles or dubbing, cost each language and review the translated video scene by scene.
- AI Voice Over for Videos: Can a Production Company Use It for Clients? — AI voice over suits client videos when the voice is licensed or consented, the client agrees in writing and platform labels are followed. The steps.
- AI Tools and AI Development: The Complete 2026 Guide — the AI hub, including every tutorial in the AI-for-business series.
Sources: Filmustage's pricing page (rechecked 4 November 2026). The 90-second script, shot counts, set-up times and schedule are illustrative planning assumptions written for this tutorial, not industry standards.