How AI Is Actually Changing Pre-Production (Not Just VFX)

The AI Conversation Is Stuck on the Wrong Stage

Most of the public conversation about AI in production is about VFX, generated footage, or de-aging actors. That's the flashy end of it, and it's also not where most productions are actually feeling the impact day to day. The bigger, quieter shift is happening earlier, in pre-production, where a huge amount of time gets spent on tasks that are repetitive, research-heavy, or just tedious enough that nobody enjoys doing them.

That's the part worth paying attention to, because it's already changing how planning gets done on smaller and mid-size productions that don't have a department for everything.

Where AI Is Quietly Useful Right Now

Script breakdowns are a good example. Pulling every prop, location, cast member, and special requirement out of a script used to take a first AD or a coordinator hours of careful reading and tagging. AI tools can now do a rough first pass of that breakdown in minutes, flagging scenes, props, and characters so a human can review and correct it instead of starting from a blank page.

Budget drafts work similarly. Early-stage budgets are often built by adapting a template from a previous job, and AI tools that have access to past budget structures can generate a reasonable first draft based on scope, location, and crew size, which a producer then adjusts rather than building from scratch.

Research tasks, like comparing locations, checking permit requirements for a city, or summarizing what gear a specific look typically requires, are also a strong fit, since AI is good at compressing scattered information into something usable quickly.

Task	Where AI Genuinely Helps	Where It Still Falls Short
Script breakdown	First pass tagging of props, locations, cast	Catching tone, subtext, and creative intent
Budget drafting	Fast first draft based on scope and past jobs	Real negotiation and vendor relationships
Location research	Compiling permit rules and comparable options	Knowing how a space actually feels to shoot in
Scheduling	Generating a draft order based on availability	Reading crew fatigue and on-set realities

What AI Still Can't Do

Every one of those examples has a ceiling, and the ceiling is judgment. AI can summarize permit requirements for a location, but it can't tell you that the neighborhood gets too noisy after 3pm because a school lets out nearby. It can generate a draft schedule, but it can't sense that a director needs a slower morning because the previous day ran long and the cast is fried. It can draft a budget, but it can't read a vendor relationship and know when a favor is worth asking for.

This is the part that gets lost in a lot of AI discourse. The tools are genuinely good at compressing information and producing a starting point. They are not good at the parts of production that depend on relationships, instinct, and being physically present for how a day actually unfolds. Anyone treating AI output as a finished decision rather than a draft is setting themselves up for a bad surprise on shoot day.

This Looks Different Depending on Budget Level

A studio feature with a full pre-production team already has people dedicated to script breakdowns, budgeting, and scheduling, so AI tools mostly speed up work that was already being done by specialists. The bigger impact shows up on smaller productions, where one coordinator or producer is often doing all of those jobs at once with limited time. For a two-person production office juggling a dozen pre-production tasks in a week, shaving hours off a script breakdown or a first-draft schedule isn't a nice-to-have, it changes what's actually possible to take on.

That's also where the judgment gap matters most. A studio production has enough people around to catch an AI tool's mistakes before they become a real problem. A smaller team relying on a single producer to both generate and review AI output has less of a safety net, which is exactly why treating these tools as a draft generator rather than a final answer matters even more at that scale.

What to Actually Look for in a Tool

Not all AI tools billed as production software are doing the same thing under the hood, and it's worth knowing what to evaluate before adopting one. A few questions tend to separate tools that save real time from ones that just look impressive in a demo.

Does it work with the documents you already have, like an actual script PDF or a real call sheet template, or does it require you to reformat everything to fit the tool first. Does it show its reasoning or sourcing, so you can verify a permit requirement or a budget number instead of taking it on faith. And does it get faster and more accurate the more you use it on your specific kind of production, or does every job start from zero again.

Tools that fail those questions tend to create more cleanup work than they save. Tools that pass them tend to disappear into the workflow in a good way, where you stop noticing the tool and just notice that the tedious part of the job takes less time than it used to.

The Real Shift Is Time, Not Magic

The honest version of this story isn't that AI is replacing pre-production work. It's that the tedious first sixty percent of certain tasks, the part that's mostly information gathering and first-draft generation, is getting faster. That leaves more time for the part that actually requires a person: reviewing the breakdown for what the software missed, adjusting the schedule around the things only a human would catch, and making the calls that come from actually knowing the project.

That's a meaningful change even without the dramatic framing. Less time spent on a first draft means more time spent on the decisions that actually move a project forward, and that's worth paying attention to even if it doesn't make for an exciting headline.

---