INSIDE THIS AI SKILL
The craft behind its decisions.
These are the principles and procedures the skill applies. You can understand the reasoning without having to perform every production task yourself.
Beats record what changes and what it costs.
A beat that says "she confronts him" is a scene description. A beat that says what is different afterwards, and what the change cost her, is a plan.
01 / Beat sheet
What a beat has to contain
- The hook
- What is on screen in the first three seconds, and why someone with a thumb on the screen does not use it.
- The change
- What is different by the end. An episode that changes nothing has no ending available to it.
- The cost
- What the change cost, and to whom. Escalation without cost is why revenge seasons collapse around episode twenty.
- The knowledge shift
- Who now knows what. This is the state the next episode inherits and the one most likely to be contradicted later.
- The ending
- The unresolved choice, named. Not "cliffhanger" - the actual decision the audience is left holding.
02 / Beat sheet
Writing the arc
- 01Write the season question in one sentence
If it takes two, the season has two arcs and one of them is going to lose.
- 02Write the final beat first
Where this ends, and what the answer to the question turns out to be. Everything before it is now a route rather than a search.
- 03Place the pilot arc endings
Five endings, each an unresolved choice, each larger than the last. This is the part the audience will actually judge.
- 04Fill in what changes between them
Work backwards from each ending to the change that makes it inevitable.
- 05Name the cost of every change
If a beat has no cost, the escalation is fake and the audience will feel it before they can name it.
- 06Check the knowledge timeline end to end
Read only the knowledge shifts, in order. Contradictions are obvious in this view and invisible in a script.
- 07Budget the beats against the format
Three principals, four locations, 60 to 120 seconds. A beat that needs a crowd scene in a new city is a beat that needs rewriting.
03 / Beat sheet
Why the first five carry the weight
The pilot arc is not a sample of the season, it is the test of it. Five episodes have to establish a world, a cast, a question, and a reason to return, and then survive contact with an audience before anyone authorizes twenty to sixty more.
Beat-sheeting the whole season and then producing five is the right order, because the five have to plant things the thirty will pay off. Beat-sheeting only the five produces a pilot that works and a season with nowhere to go.
Questions about this skill
Do I need to know how to structure a season?
No. Tell the AI Director your premise and episode count. The Season Beat Sheet Skill proposes the progression, turning points and episode endings, with dependencies between reveals and payoffs. You provide the intention and can inspect the proposed arc without having to plan every beat from scratch.
How detailed should a beat be?
Three or four lines. Long enough to state the hook, the change, the cost and the ending; short enough that the whole season can be read in one sitting, which is the only way to catch a contradiction across thirty episodes.
Should I beat-sheet all sixty episodes?
Beat the arc, in less detail toward the end. What matters is that the pilot arc plants what the season pays off. Detailed beats for episode fifty-two written before episode one is published are usually wasted work.
What if the audience data says to change direction?
Then you revise the beat sheet, which is why it exists as a document rather than as a commitment. Performance data informs the next production. What it does not do is silently rewrite approved canon.
Can the beat sheet be generated?
A draft can be proposed from a premise, and reviewing a draft is often faster than facing a blank page. The AI Director can develop the arc using this skill. Your brief, creative canon and selected review mode govern the story decisions, and rendering still requires scoped cost approval.
How does this differ for a branded campaign?
Six to ten beats instead of a season, a campaign metric instead of a retention gate, and product moments and substantiated claims placed as beats in their own right so they are planned rather than inserted.