Put them in the same room and you get production support that didn’t exist before. That sentence is the whole company, so the rest of this page is what it actually commits us to.
And yet the paperwork is where productions are won and lost. Five things are true on almost every job we’ve worked on.
The budget decides what gets shot, who gets hired and whether anyone gets paid on time. It is usually a file on a laptop, built from a previous file on the same laptop, understood in full by exactly one person. When that person is on another job, the knowledge is on another job too.
Estimated-versus-actual is reconciled after the fact on most productions below studio scale. By the time the number is visible the money is spent. An overrun you can see on day four is a decision you get to make; the same overrun found at wrap is a post-mortem.
What an Alexa 35 actually rents for in Mumbai is a different number from what it rents for in Manchester, and neither is published. That knowledge lives in the heads of people who have booked it recently. It is the most valuable thing a line producer owns and the least written down.
The 1st AD who should be solving tomorrow’s schedule is retyping a crew list. The line producer who should be renegotiating with a vendor is reformatting a call sheet at midnight. The work is necessary. Their doing it is not.
The same thirty questions get answered again. The same vendors get re-sourced. Nothing a production learns is carried into the next one, because there is no place for it to be carried to.
We are building the production layer that holds what a production knows — its rates, its vendors, its crew, what it estimated and what it actually spent — and puts a creative who has done the job in front of it. Not a tool you buy and are left alone with. Someone who deploys, with a system behind them.
If it works, a producer spends their evening on the film instead of on the call sheet, and the thing the production learned on Tuesday is still there in March.
The judgement of a good line producer is the scarce thing and it cannot be automated. The four hours a week they lose to formatting can be. We automate the second thing so there is more of the first, and we don’t confuse the two.
A creative reviews every output before it reaches you. Not as a compliance step — because a budget is an argument about what a film should be, and that argument belongs to a person. The AI does the work nobody wanted to do anyway.
Every line carries a confidence marker. Answers from the agentic layer show their sources. If you can’t see where a number came from you should not put your name on it, and neither should we.
Messaging forty crew, issuing a contract, locking a budget. These are staged, previewed and approved before they happen — enforced in the architecture rather than in a policy document. A rule that depends on everyone remembering it is not a rule.
A general model knows what a film costs. It does not know what your grip package costs from your vendor in your city this month. We build on market-specific data and on your own last three productions, because that is the difference between a plausible budget and a usable one.
What came in under. What overran. Which line the producer cut, and which vendor delivered. Every production that runs through MARK sharpens the estimate on the next one — and that data is anonymised and aggregated, never a client’s budget shown to another client.
The most personal is the most creative. We think that’s true of production too — the specific, unglamorous knowledge of how your unit actually works is worth more than any general model of how films get made.
Written on set, not in a deck