The thing that actually decides whether this works
It is not the tool. It is whether making a video is a decision you take every week, or a routine you already have.
A decision costs energy. Every single time, you have to work out what to make, remember which prompt worked, find the voice you liked, and rediscover the settings. That is why most people make three videos and stop.
A routine costs almost nothing, because the thinking was done once. This lesson is about doing that thinking once.
In plain English
- Batching:
- Doing one type of task for several videos in a row, rather than making each video start to finish.
- Content calendar:
- A simple list of what you plan to publish and when. A spreadsheet is a content calendar.
- Master prompt:
- A prompt you saved because it worked, ready to reuse with the details swapped out.
- Naming convention:
- A consistent way of naming files so you can find last month's video without opening nine folders.
Where your workflow lives
Not here. LearnAI has no accounts and stores nothing about you: your progress sits in your own browser and that is the whole system. So the checklist below belongs somewhere you own. A note on your phone, a document, a pinned message to yourself. Anywhere you will actually look on a Tuesday morning.
Copy it out now rather than promising yourself you will come back for it.
Batch by task, not by video
The instinct is to take video one from idea to published, then start video two. That is the slow way, because every switch between writing, generating and editing costs you a restart.
Batch instead. Script four videos in one sitting while your writing brain is warm. Generate all the footage in the next. Do all the voiceovers together, in the same voice, so they sound like a series rather than four strangers. Publish on a schedule.
Four videos batched takes noticeably less than half the time of four videos made one at a time, and they come out looking related, which is most of what people mean by consistency.
The weekly routine
Ninety minutes, once a week, split into four blocks. Plan the next four ideas. Write and read aloud. Generate and assemble. Voice, label, publish, done.
Pick the same slot every week and defend it like a meeting, because a routine that floats is a routine that gets skipped. Start the session by opening last week's plan rather than a blank page: half of what you need is already written, and reusing it is the point.
Give your files boring, sortable names while you are at it. Something like date, then topic, then version, means that in four months you can find the shot of the dough in about ten seconds instead of opening nine folders and giving up.
MY VIDEO CHECKLIST
BEFORE (10 min)
[ ] One idea, in one sentence. If it needs two, that is two videos.
[ ] Who is watching, and what do they already know?
[ ] Where does it get published, and what shape is the frame?
[ ] Target length, and the word budget (150 words per minute).
SCRIPT (20 min)
[ ] Hook in the first line. No throat-clearing.
[ ] One idea per sentence, nothing over 20 words.
[ ] Payoff at the end delivers what the hook promised.
[ ] One call to action. Only one.
[ ] Read it aloud into a voice memo. Fix every stumble.
[ ] Word count is within budget.
PICTURES (25 min)
[ ] Two-column plan: what is said, what is seen.
[ ] Shot prompts include subject, action, setting, light, camera, length.
[ ] Every shot ends with: no faces, no on-screen text, no logos.
[ ] Anything needing my face, my premises or my exact product:
film it on my phone instead.
[ ] Same light and time of day across all shots.
VOICE (20 min)
[ ] Sentences broken, numbers and web addresses written as spoken.
[ ] Important word at the end of a short sentence.
[ ] Voice tested on 20 words from the middle, not the opening.
[ ] Listened to the whole thing with eyes off the screen.
BEFORE PUBLISHING (15 min)
[ ] Every voice used has written permission from its owner.
[ ] Music and footage are licensed for this use, commercially.
[ ] AI label set using the platform's own setting.
[ ] Plain disclosure line in the description.
[ ] Captions on. Most people watch with the sound off.
[ ] Filename follows my naming convention.
[ ] Saved: the prompts that worked, for next time.
-
IDEAS
Give me 8 video ideas for [audience] about [topic].
Each must be one idea only, provable in under 60 seconds,
and something I can film or generate without a specific real person.
-
SCRIPT
Write a [45] second script (about [110] words) for [audience].
Facts to use: [yours]. Open with the most surprising one.
One idea per sentence, nothing over 20 words, spoken English.
-
TIGHTEN
Read this aloud in your head and list every place a speaker
would stumble or run out of breath. Quote the line, say why.
-
SHOTS
Turn this script into a two-column shot list: spoken line,
seconds, and what is on screen. Then write a shot prompt for each row.
-
VOICE PREP
Break long sentences, add breath breaks, and rewrite numbers,
acronyms and web addresses exactly as they should be pronounced.
-
FINAL CHECK
List everything in this video that needs permission or a licence,
and the one question I must answer to confirm each. Do not reassure me.
Here are the four videos I published this month, with their view counts
and how long people watched: [paste].
- What do the three best performing ones have in common?
- What do the weakest have in common?
- Give me one specific change to try next month. One only.
Be blunt. Do not tell me to post more often.
โ Weak prompt
Prompt
Make me a content plan for my business.
Output
Week 1: Introduce your brand. Week 2: Share your story. Week 3: Behind the scenes. Week 4: Customer testimonials. Post consistently and engage with your audience.
This is the same plan the assistant gives everyone, because it was given nothing to work with. Not one of these is a video you could actually start making tomorrow morning.
โ Good prompt
Prompt
I run a sourdough bakery on Bridge Street. My customers are locals aged 30 to 60 who buy bread weekly and mostly do not know why sourdough takes 18 hours. Give me 8 video ideas, each one single idea, provable in under 60 seconds, filmable with a phone or generated without showing a person's face. For each, give the hook line and the payoff line.
Output
Idea 1. Hook: your supermarket loaf took three hours, ours takes eighteen. Payoff: the extra fifteen hours is what makes it easier to digest. Idea 2. Hook: this jar is older than my daughter...
Real audience, real constraints, real facts. Every idea arrives with a hook and a payoff, so the plan is already half a script rather than a list of themes.
Video one will not be your best and it is not supposed to be. Publish it anyway. The people who get good at this are the ones who made twelve mediocre videos, not the ones still perfecting the first.
Checkpoint
Keep your checklist and prompt pack somewhere you own, batch by task rather than by video, and repeat the routine weekly so making a video stops being a decision.
Finishing the course
Pass the quiz below and the Reel Maker badge is yours. It is a small thing, but it marks a real change: you now know how to write for the ear, plan a video that holds together, generate what a machine is genuinely good at, make a voice sound like a person, and stay on the right side of consent and disclosure.
The rest is Tuesday mornings.
๐ Quiz
Question 1 of 4Why does batching by task beat making each video start to finish?