How to sell AI-made promo videos to clients (and what to charge)

If you can now produce a decent promo video for $69 — or for nothing at all, if you can live with a watermark on it — can you invoice a client two thousand dollars for it?

The instinct says no. I'd take that instinct seriously rather than argue it away, because it's pointing at something real: there is a version of this business that is just arbitrage, and it deserves the discomfort. There's also a version that doesn't, and the difference between them is entirely in what you do around the render.

For most of my career the thing I actually sold was labour. Hours in After Effects, priced roughly by how many of them a job would eat. I got very good at estimating those hours, and that's the skill currently losing its value — which is why the people most confused about what to charge now are often the ones who were best at it.

The client was never buying minutes of animation

Nobody has ever thanked me for how long something took. They thanked me when a homepage stopped needing a paragraph of explanation, or when the launch post finally had something in it worth clicking.

That's the thing being bought. A founder whose landing page isn't working isn't shopping for 45 seconds of motion graphics; they're trying to stop people leaving before they understand what the product is. If the video fixes that, the value to them doesn't change based on which software rendered the frames. And if it doesn't fix it, three weeks of handmade keyframes don't earn the invoice either. I've delivered work I was genuinely proud of that moved nothing for the client, and I still billed for every hour of it, because hours were what I'd sold.

So the honest question isn't "how much did this cost me to make." It's "what did I take off this person's plate, and what did I add."

The gap has to be filled with actual work

If your process is: paste URL, press render, email the MP4 — then yes, that's arbitrage, and it has a short shelf life. Your client can paste a URL too. Sooner or later someone will point that out to them.

Here's what sits in the gap when the job is done properly.

It starts with a brief you take seriously. Not "what colour do you want" but what they actually sell, who inside the buying company says no, and what the thing is that people misunderstand about the product. Most freelancers skip this because it feels like unpaid time, and it's the half-hour that decides everything downstream.

Then sign up and use the product. Poke around, hit the empty state, find the one screen where it becomes obvious what this is for. You cannot pick good screenshots for software you've never opened, and to my mind screenshot choice is now most of the craft.

The script needs rewriting into the client's language. A generator will hand you something competent and slightly generic. The client's customers have their own words — the noun they use for the thing, the verb they use for the action. Swapping those in takes minutes and makes the video sound like it came from inside the company.

The screens are their own job: which three views tell the story, at what zoom, with what data on them. Empty tables and Lorem Ipsum have sunk more of my early work than any animation problem did. Populating a demo account with plausible-looking data is unglamorous, and it's the step that separates your version from the client's version.

Revisions are mostly translation. "Make it pop" is not a change request; your job is to turn it into one, propose it, and get agreement before you touch anything.

And the handover is a package, not a file. Horizontal and vertical, burned-in captions for muted autoplay, a thumbnail, a still frame for the press kit, sensible filenames, and two sentences on where to embed it and why above the fold. In my experience that's the part clients bring up months later, more often than anything about the animation.

That is a service. The render is one step inside it, and it stopped being the expensive step.

Pricing follows the result, not the runtime

Two videos can be the same length and the same difficulty and be worth very different money. A landing-page video for a funded B2B tool sits in front of every trial signup for as long as the page stays up, which is usually a long time. A video for someone's personal blog does not. Charging both the same gets it wrong in both directions.

Three questions get me to a number: where will this live, how long will it live, and what happens to their business if it works. (What the work costs on the other routes is useful context when a client asks why you are not the cheapest.) A homepage hero for a product with real revenue is a four-figure job when I quote it. That's my number, not a market rate — I can't source one of those for you, and marketplace listings will show you plenty of cheaper, some of it fine.

Then sell a package, not an hourly rate. Something like: discovery call, script written and approved, one finished video, two aspect ratios, captions, thumbnail, two rounds of revisions. Everything named, everything bounded. A third revision round has a price on the same page as the first two, and you say it warmly and without apology when it comes up.

Scope creep, on the jobs where it's happened to me, hasn't usually arrived as extra revisions. It arrived as a fourth stakeholder appearing in week two with opinions nobody asked for. One point of contact, feedback consolidated into a single document, written into the agreement. That clause has done more for my week than any tool I've ever bought.

Say that you use AI

Yes, tell them. Not because of a rule, but because the alternative is a business whose stability depends on a client never finding out, and that's a bad thing to build on.

You don't have to lead with it — nobody opens a pitch with their toolchain. But put one plain line in the proposal along the lines of: production is AI-assisted; script, direction, screen selection and final edit are mine. When someone asks directly, answer directly.

Two questions come up when I've had this conversation, and both are answerable. Can they use it commercially — check your tool's licence terms and be able to quote them. And is a human accountable if it's wrong — that's you, in writing. Some organisations also have policies that rule out AI-generated assets entirely. Far better to find that out in week one than after delivery.

The first three jobs, and the three ways this goes wrong

You have no portfolio, so make one. Pick three real products you admire, make a promo for each, and label them clearly as self-directed work you weren't hired for. Passing spec work off as client work is the kind of shortcut you don't get to undo.

Narrow the offer, and know what you are competing against — agency, freelancer and AI each fail differently. "Promo videos" competes with everyone. "Landing-page videos for B2B SaaS, script to delivery in a week" competes with a much smaller field and lets you charge like a specialist. And don't try to win the cheapest-listing race; that market prices output, and output is what just got cheap.

On turnaround, promise a first cut in a few business days rather than something-by-tomorrow. Same-day promises turn you back into a render button.

Then the three things I'd expect to hit you. Someone will say "I could just do this myself" — and the honest answer is yes, you could, here's what you'd be taking on, and some of those people genuinely should. They weren't your client. Someone will have no brand guidelines, which quietly makes you the brand decision-maker; pull colours and type from their site, get them approved in writing, and treat that as scoped work. And someone will revise forever, which in my experience almost always traces back to a script nobody formally approved. Written script sign-off before any render removes most of it.

For transparency about my own end of this: I build PromoHyper. The first video on an account is free — one per account, watermarked, no revision, no card — and everything up to the final render is free either way, so you can build the storyboard, show a client the watermarked cut, and only pay once they've said yes. Buying a plan then clears the mark from that same video rather than spending a credit on it. But the prices on the pricing page assume you're making videos for your own product: if you're making them for clients, or shipping any real volume of them, that pricing stops making sense and I'd rather quote the work properly. Tell me what you're doing — canberk@promohyper.com.

Which is the part that decides whether any of this holds up. The brief, the screens, the handover — that's what the invoice is for. Take those away and you're reselling a render, and that's something your client can do without you.