Can an AI UGC Agency Reach $100K/Month? Build the Capacity Model

"$100K/month is easy because AI UGC retainers are $20K–$35K" is not a business model.

It assumes:

buyers accept the price;
scope is controlled;
clients renew;
approvals work;
results justify the retainer;
delivery cost stays low;
the founder can sell and operate simultaneously.

The revenue arithmetic is easy.

The capacity model is not.

Choose the revenue shape

Five clients at $20K

Ask your agent

5 × $20,000 = $100,000 MRR

Concentration risk:

one churn = 20% revenue loss.

Ten clients at $10K

Ask your agent

10 × $10,000 = $100,000 MRR

More account management and approval surfaces.

Twenty clients at $5K

Ask your agent

20 × $5,000 = $100,000 MRR

Likely a standardized service or software-assisted model.

Platform + service

Ask your agent

software revenue + implementation + managed creative = $100,000

Potentially more durable, operationally more complex.

Pick one. Pricing without scope means nothing.

Define the deliverable

Bad:

Unlimited viral AI UGC.

Useful:

Ask your agent

monthlyScope: researchBriefs: 4 approvedConcepts: 12 sourceFrames: 12 approvedVideoAssets: 20 platformCuts: 40 revisionRounds: 2 performanceReviews: 2 clientProvides: - current product access - approved claims - brand assets - review within 2 business days excluded: - publishing - community management - paid media buying - unlicensed creator likeness - guarantee of performance

The contract sells approved outcomes and a process, not generated file count.

Model delivery capacity

Per client, per month:

Work	Quantity	Minutes each	Hours
Research/strategy briefs	4	90	6
Concept review	12	15	3
Frame review	24 candidates	5	2
Video review	40 candidates	6	4
Client revisions	20	8	2.7
Reporting/call	2	60	2
Account/admin	—	—	3
Total	—	—	22.7

Ten clients:

Ask your agent

227 focused hours/month

That excludes:

sales;
finance;
hiring;
platform failures;
escalations;
product work.

One founder is not delivering that well.

Identify the scarce role

Generation is rarely the scarce role.

Possible bottlenecks:

creative strategist;
rights/claims reviewer;
frame approver;
editor;
client success;
paid-media analyst;
founder.

Capacity:

Ask your agent

client capacity = focused delivery hours available ÷ hours required per client

If a senior creative lead has 100 focused production hours and each client needs 12:

Ask your agent

8.3 clients theoretical

Plan below theoretical capacity.

Calculate gross margin honestly

Revenue:

Ask your agent

$100,000

COGS may include:

strategy labor;
production labor;
creator licensing;
model/API;
editing;
storage;
client success;
contractors;
quality/compliance review;
directly attributable software.

Example:

Ask your agent

delivery payroll/contractors: $35,000 creator rights: $8,000 generation/media infra: $5,000 direct client software: $4,000 total COGS: $52,000 gross profit: $48,000 gross margin: 48%

Then operating expenses:

founder;
sales;
legal;
accounting;
general software;
insurance;
office;
bad debt.

$100K revenue is not $100K income.

Include failed production

If scope promises 20 approved assets:

Ask your agent

candidates required = approved assets ÷ final approval rate

At 60%:

Ask your agent

20 ÷ 0.60 = 33.3 candidates

Then add source-frame approval and concept approval.

Price the funnel, not the final file.

For each client:

Ask your agent

concepts → approved briefs → approved frames → approved videos → deployed assets

Low approval can destroy margin even when API cost is tiny.

Design scope-change rules

Defect:

The generated hand is broken.

Included retry.

Changed brief:

We approved the script, but now legal wants a different positioning.

New work.

Contract should define:

included revisions;
defect correction;
changed product facts;
new format;
new creator;
rush;
paused approvals;
cancellation;
carryover;
expired trend work.

Without it, every client decision becomes agency COGS.

Model client approval latency

If client review takes seven days, weekly production will queue.

Set:

approval SLA;
maximum pending queue;
consolidated feedback owner;
stale-approval rule;
no generation beyond WIP limit.

Example:

When ten assets await client review, new motion generation pauses. Strategy may continue up to the agreed brief limit.

This protects both margin and quality.

Model sales capacity

Let:

target new clients = 2/month;
close rate from qualified opportunity = 20%;
show rate = 80%;
opportunity rate from qualified meeting = 50%.

Required qualified meetings:

Ask your agent

2 ÷ 0.20 ÷ 0.80 ÷ 0.50 = 25/month

Then calculate how many leads create 25 qualified meetings.

Do not assume a founder can:

run 25 meetings;
close;
onboard;
approve creative;
handle clients;
build the product.

Assign ownership.

Model churn

At ten clients:

Ask your agent

monthly churn of 10% = one client lost/month

To grow by two net:

Ask your agent

three new clients required

Track:

logo churn;
revenue churn;
expansion;
gross retention;
reason;
cohort;
time to first approved asset;
time to first business result.

The best retention lever may be better onboarding, not more videos.

Sell a learning system

Weak offer:

100 AI UGC videos every month.

Stronger:

We turn approved product evidence into a measured creative testing system: 12 concepts, 20 approved assets, lineage, and two decision reviews.

The client should know:

what was tested;
what changed;
what passed;
what it cost;
what happened;
what happens next.

That is harder to commoditize than file volume.

Keep outcome claims bounded

Do not guarantee:

views;
ROAS;
revenue;
virality;
platform safety.

You can commit to:

process;
approved scope;
review timing;
reporting;
rights;
quality;
experiment design;
transparent cost.

Performance depends on product, offer, media, market, and measurement.

Staffing sequence

0–3 clients

founder sells and owns strategy;
one strong production operator;
narrow offer;
document every stage.

4–7 clients

dedicated client success;
separate creative review from generation;
formal WIP limits;
weekly margin by client.

8–12 clients

creative lead;
production ops;
rights/compliance support;
account ownership;
backup coverage;
standardized reporting.

Beyond

productized workflows;
specialization;
portfolio allocation;
sales capacity;
management;
stronger data systems.

The exact sequence depends on scope.

A $100K example

Ask your agent

8 clients × $12,500 = $100,000

Per client:

12 approved concepts;
20 approved assets;
40 platform exports;
two strategy reviews;
no publishing or media buying.

Team:

founder/strategy;
creative lead;
two production operators;
client success;
part-time rights/compliance;
contractors for editing peaks.

Target:

client-level gross margin visible;
review queue below limit;
time to first approved asset under SLA;
no unsupported claims;
no unlicensed likeness;
outcome reporting connected to creative lineage.

This is plausible only if observed delivery hours support it.

The counterargument

Maybe an agency should not aim for $100K/month.

If higher revenue requires:

bad-fit clients;
fragile concentration;
fake claims;
burned-out reviewers;
lower gross margin;
unlimited scope;
account-evasion tactics;

then $60K with clean operations is a better business.

Revenue is not the objective function.

What Ghostfeed changes

Ghostfeed can reduce:

source-to-variant coordination;
frame handoff;
repetitive editing;
asset retrieval;
creator-format reproduction;
approval ambiguity.

That can increase the number of clients a good team serves.

It should not be used to promise unlimited output.

No filming, no editing, no prompting removes work from the customer experience. The agency still owns research, judgment, rights, and outcomes.

For starting the offer, read How to start an AI UGC agency without selling slop. For delivery economics, use AI UGC unit economics.