What a Platform Should Do Between Signing and Script Lock on an AI Vertical Drama Commission
The agreement is signed. The kickoff call went well. Then the commissioning team goes quiet for a stretch, the production partner starts writing, and the first time anyone on the platform side reads the full script is the week it is meant to lock. Notes arrive all at once. Some contradict each other. A few reopen decisions everyone assumed were settled at pitch stage, and the partner is left choosing between rewriting under pressure and locking a script nobody fully owns.
This is one of the most common failure points on a commissioned vertical drama, and it rarely looks like a failure at the time. It looks like a busy commissioning team and a partner getting on with the work. The cost appears later, in a generation phase carrying unresolved story questions, in retakes driven by script changes rather than shot quality, and in an episode one that tries to serve three different visions of the series at once.
The period between signing and script lock is the cheapest moment in the whole commission to make a decision. Nothing has been generated. No character references have been approved. Every change is still a change to words on a page. A platform that uses this window deliberately gets a script it recognises as its own, and a production partner that can move into generation with confidence. What follows is a six step framework for doing exactly that.
1. Name One Owner and One Decision Path
Why a single owner matters before anything else
The first task after signing is not creative. It is organisational. Someone on the platform side needs to be named as the owner of the commission, with the authority to give final notes and approve the script. Without that person, notes arrive from whoever has time to read a draft, and the production partner has no way of knowing which notes carry weight. A head of content, a commissioning editor and a regional lead can all have valid opinions on the same scene. Only one of them should be the person who resolves them.
What the decision path should contain
Beyond the owner, the platform should write down who else reads the script, at which stage, and what kind of input each person is expected to give. A legal or standards reviewer reads for policy. A marketing lead reads for positioning. A data lead might read episode one for hook strength. Each of these is useful, but only if the owner collects and reconciles their input before it reaches the partner. The partner should receive one consolidated set of notes per draft, not five parallel threads.
This is also the moment to agree how quickly notes will be returned. A partner can plan around a predictable turnaround far better than around a fast but irregular one. A simple rule, such as consolidated notes within a fixed number of working days after each draft, removes most of the friction this phase usually generates.
2. Confirm the Creative Thesis Before Anyone Writes Episode One
Restate the pitch in a single page
Pitches compress. By the time a series is signed, the commissioning team and the production partner often hold slightly different versions of it in their heads. One remembers the revenge arc. The other remembers the romance. Both are right, and the difference will surface in episode three if it is not surfaced now.
The fix is a one page creative thesis, agreed before scripting begins. It names the genre and subgenre, the central relationship, the protagonist and what she wants, the antagonist and what he controls, the tone, and the emotional promise the series makes to a viewer who keeps paying past the free episodes. It should also name the two or three moments the platform considers essential, the beats that made the series worth commissioning in the first place.
Agree the arc shape, not just the premise
A vertical drama series lives or dies on its arc across sixty to one hundred short episodes. The thesis page should therefore include a rough arc: where the first paywall moment falls, where the midpoint reversal lands, and how the finale resolves. None of this needs to be detailed. It needs to be shared. A platform that signs off the arc shape at this stage will rarely be surprised by the script, because the script will be an execution of something it already approved.
Keep the thesis page alive
The thesis page is not a one time exercise. It should sit at the top of every draft review, and every consolidated note should be checked against it before it is sent. A note that pulls the series away from the thesis is not necessarily wrong. Sometimes a draft reveals that the original idea was weaker than a new direction the writing has found. But that is a decision for the owner to make consciously, with the thesis page updated to match, rather than a drift that happens one note at a time. A thesis that changes on purpose is healthy. A thesis that changes by accident is how a platform ends up with a series that no longer resembles what it commissioned.
3. Settle Platform Requirements While They Are Still Cheap
Content policy, rating and market rules
Every platform has rules that shape what can appear on screen. Some are content policy. Some come from app store ratings. Some are specific to the markets the series will run in. These rules are easy to apply to an outline and expensive to apply to a finished episode. The window before script lock is where they belong.
The commissioning team should send the partner its current content policy, any market specific restrictions, and a short list of recent notes it has given other series on policy grounds. That last item is often the most useful. It shows the partner how the policy is actually interpreted, rather than how it reads on paper.
Technical and delivery requirements
The same applies to delivery. Episode length range, aspect ratio confirmation, subtitle and caption requirements, title card conventions, and whether the platform wants clean versions without burned in text. None of these change the story, but several of them change how a scene is written. A series that will be dubbed into four languages benefits from dialogue that avoids wordplay. A series that will run with on screen captions benefits from fewer overlapping lines. Writing for delivery is easier than rewriting for it.
4. Review the Script in Passes, Not in One Read
The problem with a single full read
A full script for a vertical drama can run to well over a hundred pages across all episodes. Asking one person to read it end to end and return useful notes is asking a lot, and it tends to produce notes that are heavy on the first ten episodes and thin on everything after. The back half of a series, where retention is often won or lost, gets the least attention.
A three pass structure
A more reliable approach is to review in passes, each with a different purpose.
The first pass is structural. It covers the episode by episode outline only, checking that the arc matches the thesis, that cliffhangers land at the end of every episode, and that the paywall moment falls where the platform wants it. Notes at this stage are about order and shape.
The second pass is the hook pass. It covers the first five to ten episodes in full script form, because these episodes decide whether a viewer stays. The platform should read them as a viewer would, on a phone, and ask whether each episode earns the next tap.
The third pass is the continuity and policy pass across the remaining episodes. By this point the structure is agreed and the opening is strong. What remains is to check that characters behave consistently, that nothing drifts against content policy, and that the finale pays off what the opening promised.
Splitting the review this way means each pass is short enough to be done properly, and each set of notes is focused enough for the partner to act on cleanly.
5. Approve the Visual Foundations Alongside the Script
Why visuals belong in the pre lock window
On an AI native production, the visual foundations of a series are built from the script: character descriptions, wardrobe, key locations, the overall look. If those foundations are approved only after script lock, any visual decision that conflicts with the script forces either a script change or a compromise on screen. Running the two in parallel avoids that.
What the platform should see
Before lock, the platform should expect to see and approve a small set of visual references from the production partner. Typically this means character look references for the main cast, a reference for each recurring location, and a short tone reference showing how the series will feel in colour and light. The platform does not need to approve every detail. It needs to confirm that the protagonist looks like the protagonist it commissioned, and that the world of the series matches the thesis.
This is also where casting decisions in the broad sense get made. Age, styling, the visual signal of wealth or status, the contrast between protagonist and antagonist. These are commercial decisions as much as creative ones, and they are best made by the platform while they can still shape the script, rather than after the script has fixed a description that no longer fits.
6. Define What Lock Means and How Changes Travel After It
Lock is a status, not a feeling
Many commissions treat script lock as a moment of general agreement rather than a defined status. That is where trouble starts. If lock is not defined, every later note can be framed as a small adjustment, and the script never quite stops moving. Generation then runs against a moving target.
The platform and the partner should agree, in writing, what lock means for this commission. A useful definition covers three things. The version of the script that is locked, identified by date or version number. The people who signed it off. And the process for any change after that point.
A simple change process
After lock, changes should be possible but deliberate. Borrowing from the discipline of change control used in quality managed industries, each post lock change should be requested in writing, assessed by the partner for its effect on episodes already in generation, approved by the platform owner, and recorded against the locked version. This does not need heavy tooling. A shared change log with a short entry for each request is enough.
The point of the process is not to discourage changes. It is to make the cost of each change visible before it is approved. A line of dialogue in an episode that has not yet been generated is trivial. A change to a character look in an episode that has already been generated and reviewed is not. When both sides can see that difference, decisions get better.
What to do in the final days before lock
In the last days before lock, the owner should do one final read of the thesis page against the script, confirm that every consolidated note has been addressed or consciously declined, and confirm that the visual references and the script describe the same series. Then the owner signs. The partner should leave this phase holding a locked version number, a signed off visual foundation, and a written change process. That combination is what allows generation to begin without unresolved questions underneath it.
Axis AI Studios Perspective
The pre lock window is where Axis AI Studios does a large share of its most important work on a commission, and much of it is invisible on screen. The goal is simple. Arrive at generation with a script the platform recognises as its own, visual foundations the platform has already seen, and a change process both sides agreed before it was needed.
In practice that means asking for a single owner on the platform side, working from a one page thesis, reviewing in passes, and building the visual references in parallel with the script rather than after it. It also means being direct about which decisions are cheap now and expensive later, so a commissioning team can spend its attention where it has the most effect.
None of this is complicated. It is a matter of discipline, applied early, by people who have seen what happens when it is not. Platforms that commission this way tend to get fewer surprises in delivery, cleaner episode ones, and a working relationship that carries into the next series.
For a platform preparing a commission and looking for a production partner that treats the pre lock window as part of the work, reach out at business@axisaistudios.com.
FAQ
How long should the period between signing and script lock take?
It depends on episode count, the number of reviewers on the platform side, and how quickly consolidated notes are returned. Rather than fixing a duration, it is more useful to fix the number of drafts and the turnaround for notes on each one. A clear note cadence usually does more to keep this phase on track than a deadline does.
Can generation start before the script is fully locked?
It can, in a limited way. Character and location references can be developed in parallel with the script, and some test generations can help confirm the look. Full episode generation should wait for lock, because any script change after that point carries a cost in work that has to be redone.
Who on the platform side should own the commission?
Ideally the person with the authority to approve the script and the clearest view of why the series was commissioned. That is often a commissioning editor or head of content. What matters most is that one person consolidates notes from everyone else and gives the production partner a single, reconciled direction.
Further Reading
For teams that want a clear handover process once the script is final, the guide to handing over a locked vertical drama script covers how to keep a locked script from changing without explicit approval.
For commissioning teams working through policy and standards questions before lock, the platform content policy checklist for vertical drama scripts covers what to check and when to check it.
For platforms setting up the working relationship in the first weeks after signing, the commissioning side onboarding checklist for an AI production partner covers the documents, access and decisions a partner needs from day one.

Let's set
the new standard together.
If you're working on something, we'd like to hear about it.
