How to Run an Internal Approval Process for an AI Vertical Drama Commission

Episode fourteen is finished and sitting in a review folder. It has been there for nine days. Not because anyone rejected it, and not because anything is wrong with it. It is there because the marketing lead wants to see the thumbnail frame before sign off, the legal reviewer is on leave, and nobody inside the commissioning organisation can say out loud whether marketing is actually a required approver or merely an interested one. Meanwhile the production partner has stopped work on episodes fifteen through twenty, because continuing would mean generating twenty more episodes against a standard that might be about to change. The commission has not failed. It has simply stopped, and it stopped inside the buyer, not inside the production.

This is the most common way a first AI vertical drama commission goes wrong. Not a quality failure. Not a partner failure. An approval failure, caused by an approval process that was never designed, only inherited from whatever the organisation already used for campaigns or licensing deals. Those processes assume a small number of large decisions spread across months. A vertical drama commission is the opposite shape. It is a large number of small decisions arriving quickly, most of which are reversible, a few of which are not, and almost none of which benefit from a committee.

The fix is structural rather than cultural. You do not need people to care more or respond faster. You need fewer people in each decision, clearer evidence attached to each decision, and a written default for what happens when the window closes without an answer.

1. Understand Which Decisions Are Actually Irreversible

Start by sorting the decisions, not the people. Most commissioning organisations do the reverse. They begin by listing stakeholders, then work out what each stakeholder should see, and end up with a process where everyone sees everything and nobody owns anything.

There are only a handful of genuinely irreversible decisions in a vertical drama commission. The choice of production partner and the terms of the agreement. The series premise and the lead character design, because those propagate into every downstream asset. The production standard, meaning the visual and audio bar that the series will be held to. And final acceptance of delivery, because that triggers payment and closes the revision window.

Everything else is reversible at a cost that is usually smaller than the cost of the delay required to approve it. A scene that lands wrong can be regenerated. A wardrobe choice that reads as too formal can be corrected in the next block. An episode title can be changed after delivery. When an organisation treats reversible decisions with the same ceremony as irreversible ones, the process slows by an order of magnitude and buys almost nothing in risk reduction.

So the first act of designing the process is drawing the line. Irreversible decisions get gates. Reversible decisions get a named owner who decides alone and informs afterwards. Write that distinction down before you write anything else, because every later argument about the process will be resolved by pointing back to it.

2. Build Four Gates, Not Fourteen

Four gates cover a full commission from first conversation to accepted delivery. Adding more does not add control. It adds surface area for stalling.

Gate One: Concept and Commercial Case

This gate answers whether the organisation wants a series at all, at roughly what scale, aimed at which audience, and against which commercial outcome. It closes before any partner conversation becomes substantive. The evidence it needs is a one page premise, an audience definition, a target platform or distribution route, and a budget band rather than a budget number. The decision owner is whoever holds the content budget.

Gate one fails most often when it is skipped. A commissioning team talks to three production partners, develops a favourite, and only then goes internally for a mandate. By that point the internal conversation is no longer about whether to commission. It is about whether to overrule a colleague who has already invested weeks. That is a worse conversation, and it takes longer.

Gate Two: Partner and Agreement

This gate closes the production agreement. It carries legal review, procurement review where the organisation has a formal procurement function, and the commercial sign off. It is the slowest gate by nature and the one most worth front loading, because everything after it is blocked by it.

The evidence pack here is the agreement draft, the deliverable specification, the revision and retake terms, the rights and ownership position, and the payment schedule. Run legal and procurement in parallel rather than in sequence. Sequential review of a production agreement routinely doubles the elapsed time for no improvement in the outcome, because the two reviewers are looking at different clauses.

Gate Three: Production Standard Lock

This is the gate most organisations do not know they need, and its absence is the single largest cause of mid production stalling. Before volume generation begins, the commissioning side and the production side agree on what finished looks like. Character reference sheets. Colour and lighting treatment. Audio bar. Pacing reference. Subtitle and burn in policy. Two or three fully finished reference episodes or sequences that both sides accept as the standard.

Lock it, and every later quality conversation becomes a comparison against an agreed reference rather than an argument about taste. Skip it, and every episode becomes a fresh negotiation, which is exactly the condition that produces the nine day review folder.

Gate Four: Delivery Acceptance

The final gate. It answers whether the delivered series meets the locked standard and the contracted specification. It is a conformance check, not a creative review. Creative opinion belongs at gate three, before production. Reintroducing it at gate four is how a commission ends in a dispute.

3. Give Every Gate an Owner, a Window and a Default

A gate without a named individual is not a gate. It is a queue. Assign one person per gate, by name and not by department, who holds the decision. Other people may advise. One person decides.

Then attach a window. Gate one, three working days. Gate two, ten working days, run in parallel. Gate three, five working days. Gate four, five working days from receipt of the full delivery package. These numbers are starting points and should be tuned to the organisation, but the principle does not move: every gate has a stated duration, agreed in advance, and visible to the production partner.

The part that most organisations leave out is the default. What happens when the window closes and the owner has not responded. Silence has to mean something, and the only two options are silent approval or silent rejection. For reversible decisions, silence should mean approval, because the cost of a wrong reversible decision is lower than the cost of a stall. For the four gates, silence should mean escalation to a named second, not automatic approval and not indefinite waiting. Write both rules into the process document and tell the production partner what they are, so they know when they are allowed to keep moving.

4. Separate Deciders From Informed Parties

The approval list grows because nobody wants to tell a colleague they are not needed. So the list becomes a courtesy exercise, and courtesy is expensive when it is wired into a critical path.

Split the list explicitly. Deciders can block. Consulted parties are asked for input inside the window and their input is considered but does not block. Informed parties receive the outcome and do not receive the draft. Most organisations find that a commission with eleven people attached actually has two deciders, three consulted and six informed.

The conversation that establishes this is uncomfortable once and saves weeks thereafter. It helps to frame it around the decision rather than the person. Marketing is not being excluded from the series. Marketing is being moved from a gate that blocks generation to a gate that shapes launch, which is where marketing input actually changes the outcome.

5. Build the Evidence Pack Once and Reuse It

Approvers stall when they cannot tell what they are being asked to approve. The most reliable accelerant in the whole process is a standing evidence pack format that looks identical at every gate.

Keep it to one page plus attachments. The decision being asked for, stated as a single sentence. The owner. The window and the date it closes. What changes if the answer is yes. What changes if the answer is no. The two or three risks that a reasonable person would raise, already answered. Then the attachments, which are the actual artefacts under review.

The discipline that matters here is the pre answered risk section. When an approver reads a pack that has already surfaced the objection they were forming, they approve. When they read a pack that has not, they raise the objection, which starts a round trip, which costs three days. Commissioning teams that write good evidence packs are not better at persuasion. They are better at predicting the three questions and answering them in advance.

For a fuller view of what belongs in the pack at the agreement stage, the Axis commission checklist for businesses covers the documents and terms worth settling before signature, and it doubles as a gate two evidence list.

6. Handle Mid Production Change as Change Control, Not as Approval

Changes will arrive after gate three. A platform partner asks for a different episode length. A legal note lands on a character name. A senior stakeholder sees an early episode and wants a tonal shift.

The failure mode is treating these as fresh approvals, which sends them back through the whole gate structure and stops production while they travel. The alternative is to treat them as change control items, which is the standard discipline for altering a specification that work is already running against.

A change control item has four fields. What is being changed. What it affects, listed concretely as episodes, assets or scenes. What it costs in money and in schedule. And who is accepting that cost. It goes to one decider, not to the gate structure. Production continues on everything the change does not touch, which is usually most of the series.

This single move preserves more schedule than any other part of the process, because it stops one localised change from freezing an entire slate. It also creates an honest record. When the fourth tonal change arrives, the accumulated change log makes the cost of indecision visible in a way that no meeting ever does.

7. Give the Production Partner Visibility Into Your Own Process

Commissioning organisations tend to treat their internal approval process as internal. The production partner sees only the output, which is an answer or a silence.

That asymmetry causes defensive behaviour on the production side. A partner who cannot see where a decision sits will either stop work to avoid generating against a standard that might change, or push on and risk rework. Both are expensive, and both are caused by missing information rather than by bad judgement.

Share three things. The gate map, so the partner knows how many decisions remain and what each one covers. The current gate status, updated weekly. And the named owner of the open gate, so the partner knows whose calendar is the constraint. None of this requires exposing internal politics. It requires exposing structure.

A partner with this visibility can sequence its own work around your decision windows, front loading the parts of the series that no open gate touches. That is schedule recovered at zero cost.

8. Measure the Process After the First Commission

Run the first commission, then measure four things before the second one starts.

Total elapsed days per gate, against the window you set. Number of decisions that returned for a second round, and why. Number of change control items raised after gate three, which is a direct measure of how well gate three was run. And days of production idle time attributable to an open internal decision, which is the number that converts process cost into a figure the organisation recognises.

Most organisations discover the same two things. Gate three was under invested, and the change log after it is long as a result. And the real constraint was not the number of approvers but the absence of a default for silence. Both are cheap to fix in the second commission, and fixing them is what turns a first commission into a repeatable commissioning capability rather than a one off project that happened to work.

The budgeting side of this is worth settling in the same pass, since approval delay usually surfaces as a cost question before it surfaces as a schedule question. Idle production time has a price, and putting a figure on it is what moves approval design from a process conversation to a budget one.

Axis AI Studios Perspective

Axis AI Studios works with commissioning teams whose internal approval structures were built for something other than serialised vertical drama, and the gate map above reflects what those engagements settle into. The production side of a commission is where Axis carries responsibility, and the part of the approval process Axis controls is the part that makes the buyer side decisions easy to make: a locked production standard agreed before volume work begins, reference material that makes conformance checkable rather than debatable, a change control format that separates what a change touches from what it does not, and visible status on every open decision that holds work.

That is the recognition worth stating plainly. Commissioning teams are not slow. They are asked to make fast decisions without the evidence that would make the decision obvious, and they are asked to make them collectively when they should be making them individually. A production partner that supplies decision ready evidence at each gate removes most of the reason a commission stalls, and does so without asking the organisation to reform how it governs anything else.

Teams planning a first commission and mapping the internal approvals around it can start the conversation at business@axisaistudios.com.

FAQ

How many approvers should a vertical drama commission have?

Two deciders across the whole commission is common and workable. One holding the commercial and partner decisions, one holding the creative standard. Consulted parties can be more numerous, but only deciders should be able to block, and only one decider should hold any single gate.

What should happen if an approver misses their window?

For reversible decisions, the default should be approval, so that production continues. For the four structural gates, the default should be automatic escalation to a named second decider rather than open ended waiting. Both rules need to be written down before the commission starts, not decided during the first stall.

Can the production standard be changed after it is locked?

Yes, but through change control rather than through the gate process. The change is scoped against what it touches, costed in money and schedule, and accepted by one decider. Everything the change does not touch keeps moving, which is the whole point of handling it this way.

Further Reading

For the documents and commercial terms to have settled before an agreement reaches legal, the commission checklist for businesses covers what needs to exist before signature.

For the structural work that makes gate two move faster, the guide to structuring a commission for legal review covers how to present a production agreement so legal can assess it in one pass.

For the budget framing that sits under gate one, how to budget your first vertical drama commission covers what the numbers look like before a band becomes a figure.