"What does it cost?" is the first question every brand asks and the one almost no generative AI consultant answers in public. Here is the straight version, from someone who quotes this work every week.
The three engagement shapes and their price logic
1. Project-based work
A defined deliverable: a launch film, a batch of product ads, a UGC series. This is the easiest to price because the scope is concrete. The price is a function of:
- +Number of final assets and variants
- +Usage rights: organic only versus paid media versus perpetual
- +Revision depth: locked brief versus exploratory direction
- +Turnaround: standard versus compressed
A single polished AI product ad costs a fraction of a traditional shoot, which routinely runs five figures once crew, talent, location, and post are counted. That delta is the entire reason this market exists.
2. Retainers
Ongoing production plus the strategy layer: what to make, where to publish, what to kill. Retainers make sense once a brand wants content weekly instead of once a quarter. Pricing scales with volume and how much of the consultant's calendar you are reserving.
3. Systems builds
The high-leverage one: a one-time build of an agentic content automation engine your team runs afterward. You are not buying videos, you are buying the machine that makes videos. Priced accordingly, and worth it only if you will actually feed and operate the machine.
What moves the number up
- +Paid media usage. An ad that will carry real spend gets more iteration, more QA, and broader rights. Expect a multiple of organic-only pricing.
- +Likeness and compliance. Real people, regulated industries, and brand-safety review add cycles.
- +Consistency requirements. A one-off video is easy. A series with persistent characters, sets, and grading is a different discipline (this is exactly the frame-chaining workflow problem).
- +Speed. Compressed timelines compress everything except the price.
What moves the number down
- +A clear brief with references. Ambiguity is billed.
- +Existing brand assets: product shots, fonts, color systems.
- +Batching. Six ads scoped together cost less per ad than six separate engagements.
- +Organic-only usage to start, with an option to upgrade rights later.
Why most consultants say "rates on request"
It is not coyness. The honest cost depends on volume, rights, and system ownership, and a public rate card either overcharges small scopes or underprices serious ones. What you should demand instead of a rate card is a fast, concrete quote. My standard: tell me the brand, the goal, and the timeline, and you get a concept and a number within 48 hours.
The cost comparison that actually matters
Do not compare a generative AI consultant's quote to "doing it ourselves with a subscription." Compare it to the real alternatives:
- +A production agency: weeks of lead time, five-figure day rates, one concept
- +An in-house hire: salary, ramp time, and tool churn they will fight alone
- +Doing nothing: your competitors' feeds are not waiting
The consultant's job is to be cheaper than the first, faster than the second, and to make the third indefensible.
Questions to ask before you sign
- 01Can I see shipped work in my format? (Mine is here.)
- 02Who owns the prompts, workflows, and source files? (You should.)
- 03What happens when a better model ships mid-engagement? (The pipeline should absorb it.)
- 04What are the usage rights by channel?
- 05What does week one look like?
If the answers are crisp, the rate is rarely the problem. If they are vague, no rate is low enough.