7-lesson course · about 16 hours

Run three agent teams from one scene record, and keep every call with its owner

For Filmmakers, producers, editors and production teams coordinating a story across development, planning and post-production

Everything for Filmmakers

A self-paced capstone for production leads. Connect story, planning and post agent teams in Dots or Claude Duet, carry a late revision through all three, then test a restart.

Try lesson 1 freeSee what you’ll learn

What you’ll be able to do

By the end, you can…

Design three supervised agent teamsSet up Story and Continuity, Production Planning and Post Coordination for one short sequence. Each team gets a real deliverable, clear edges and acceptance conditions you can check. Each type of decision is mapped to the named person who owns it.
Keep every handoff in one shared recordBuild a handoff table that records the script version, accepted inputs, deliverable, owner and status at every handoff. Only accepted inputs can become a baseline, so a draft never quietly turns into the plan.
Have Dots or Claude Duet follow the approved recordConfigure your chosen agent to read only approved entries and return revision-linked impacts, department questions and outstanding post decisions. Availability, costs and safety procedures go to their owners as questions, never as made-up answers.
Carry a contradictory revision through every teamWork through the night-to-dawn rewrite that arrives after the rain wide is approved. Stop incompatible work, keep both versions, and route creative and likeness calls, such as relighting a performer's face, to the director and the performer.
Break the Factory on purpose, then restart itTest changing inputs, blocked work, missing media and a later restart. Resume only when every team holds the same facts and no accepted work has been overwritten without notice.
Leave a record another lead can pick upRefine the team instructions, run a second scene packet and write decision entries. Anyone picking up the record should be able to see the current script, selected shots, open issues, the next decision, where AI touched the picture and who signed off.

How it works

A tutor that meets you where you are

  1. Lessons pitched at your levelSeven lessons in three modules follow the Capstone in order: define the teams and handoffs, carry a revision through them, then refine the process and approve it for the next sequence. The lessons adapt to how much agent and production experience you bring.
  2. An AI tutor you can ask anythingType or speak your questions while you work. Ask why a stop fired, how to word an acceptance condition, or how to adapt the sample scenes to your own packet.
  3. Quizzes and flashcards for each lessonShort checks tied to each lesson cover handoff rules, gate placement and who owns which decision, so the distinctions stick.
  4. Progress tracking and a certificateTrack where you are across the modules and pick up a run where you left it. You receive a certificate of completion when you finish.

Curriculum

The curriculum

The course runs in three modules: connect the teams around one scene record, carry a revision through every team, then refine the process and record the decisions for the next sequence. You can work on your own three-to-four-scene packet or the sample packet for scenes 41–44.

Module 1Connect story, production and post proactive AI agent teams.2 lessons

  1. 1Define the three proactive AI agent teams
    • Define the Story and Continuity, Production Planning and Post Coordination teams, with roles and edges for each
    • Assign each team a real deliverable and write acceptance conditions that can be checked
    • Map each type of decision to the named person who owns it
    • Choose OpenAI Dots or Claude Duet for the Factory and explain the choice
  2. 2Define the department handoffs
    • Build a handoff table that records script version, accepted inputs, deliverable, owner and status for every handoff
    • Explain the difference between an accepted input and a draft, and why only accepted inputs can be a baseline
    • Configure Dots or Claude Duet to read only approved entries and return revision-linked impacts, department questions and outstanding post decisions
    • Place the performer-approval and pre-render sign-off gates in the record

Module 2Carry one revision through every proactive AI agent team.2 lessons

  1. 3Run the proactive AI agent team handoffs
    • Run a sequence packet through all three teams with every handoff recorded
    • Use accelerated checks to test work that would normally wait on departments
    • Test the Factory with changing inputs, blocked work and missing media
    • Test a later restart and confirm every team resumes from the same facts
  2. 4Resolve a contradictory revision
    • Identify which accepted work a contradictory revision breaks in each team
    • Stop incompatible work while keeping both versions in the record
    • Route creative and likeness questions, such as relighting a performer's face, to the director and the performer
    • Confirm that every team holds the same facts and that nothing accepted was silently overwritten before resuming

Module 3Approve the process for the next sequence.3 lessons

  1. 5Refine the proactive AI agent team instructions
    • Diagnose missed stops, unneeded stops and unclear questions from the run records
    • Tighten each role's boundaries, retry limits and escalation routes
    • Version the instructions in the scene record
    • Test the revised instructions against the earlier failure cases
  2. 6Run a second scene packet
    • Run a second scene packet through all three teams using the refined instructions
    • Check that an earlier decision reaches every team correctly
    • Compare stops and handoffs with the first run
    • Raise the pickup question without letting the agent schedule it
  3. 7Record the department decisions
    • Write decision entries that name the owner, revision, alternatives kept and effect on each team
    • Test whether another lead can read the current state from the record alone
    • Record where AI touched the picture and who signed off
    • Explain the consent, licensing and crew-time considerations a lead should record

Who it’s for

Made for you if…

Made for: Filmmakers, producers, editors and production teams coordinating a story across development, planning and post-production

Producers and production coordinators

You already know a script change travels through shots, schedules and the edit. This course shows you how to have agent teams carry that paperwork while scheduling, resource and release decisions stay with you and your department heads.

Editors and post supervisors

You will set up a Post Coordination team that flags missing media, keeps old notes off replacement shots and puts sign-off ahead of the render queue, because an approved take is copied into every deliverable.

Directors and filmmakers who have built single agents

You finished the Intro and Intensive, or have similar experience, and now want several bounded agents working from the same facts across departments without taking creative or likeness calls away from you and your performers.

Questions

Good to know

What do I need before I start?

You should have finished the Film AI Intro and Intensive (days 1–7) or have equivalent experience building bounded proactive agents with stops and escalation. You should also know production paperwork (breakdowns, schedules, call sheets, shot lists, continuity notes, edit handoffs) and script revision practice such as blue pages and scene numbering. You need your own access to OpenAI Dots or Anthropic Claude Duet.

Should I use Dots or Claude Duet?

Either one. The course treats them as equal options and does not benchmark one against the other. In the first lesson you choose one and write down why.

Do I need my own scenes?

No. You can use your own three-to-four-scene packet or the course's sample packet for scenes 41–44.

Will I end up with a finished film or a production-ready schedule?

No. The Capstone tests coordination across departments. It does not produce a finished film, a production-ready schedule or a real budget.

Will the agents make scheduling or creative decisions?

No. The agents raise impacts and questions. Scheduling, resource, release, creative and likeness decisions stay with the named department owners, the director and the performer. Availability, costs and safety procedures are sent to their owners as questions.

Does the course teach shot generation or upscaling?

No. Hands-on generation, finishing and upscaling are out of scope, and so is building single agents from scratch, which the Intro and Intensive cover.

Does it cover union rules, likeness rights or licensing?

It explains what a lead should record about consent, licensing and crew time. It is not legal advice on union agreements, likeness and voice rights or IP licensing.

How long does it take?

It takes about 16 hours, at your own pace.

Film AI Capstone: a supervised Factory of agent teams for sequence preparation

$397

One payment · lifetime access

  • 7 self-paced lessons in 3 modules, about 16 hours
  • A handoff table and decision record built around one versioned scene record
  • Guided tests: a contradictory revision, changing inputs, blocked work, missing media and a later restart
  • Use your own scene packet or the sample packet for scenes 41–44
  • An AI tutor you can type or talk to, plus quizzes and flashcards for each lesson
  • Progress tracking and a certificate of completion

Secure checkout by Stripe. Have a promotion code? Enter it at checkout.

Refunds: Full refund if you ask before you open your first lesson. Once you’ve started, purchases aren’t refundable, except where the law gives you that right.

Already bought it? Sign in