◆ CONSULTING PROPOSAL · APRIL 2026 ◆

YOUR IN-HOUSE
AI CREATIVE
SYSTEM.

:: A PROPOSAL FOR MEDIA MADE SIMPLE · ROOSTER + WALT ::
» MISSION BRIEF «

You spent two to three weeks trying to make AI statics work in Claude and Antigravity and couldn't get anything decent out. The pieces are there. What's missing is the cookbook — the markdown files, the agent loops, the QC patterns my team uses to ship 14 statics a week per client. This is a 3-month engagement to install it inside your agency.

3-month engagement 4 weekly calls in month 1 Slack-embedded throughout Adlib + kit drops live
▶ LET'S RUN IT
M1 is the build sprint — the rest runs on light calibration.
CHAPTER I · WHERE YOU ARE

You're not stuck on the model. You're stuck on the system around it.

From our call on April 24. Media Made Simple is brand-forward, performance-driven, already running paid social and Google for DTC brands doing $100k+/mo. The thing in the way is the AI creative pipeline. Two pieces of the call this proposal is built around:

▶ ROOSTER · ON THE BUILD

"I gave it a proper go. Tried to sprint two to three weeks trying to make that work within Claude and Antigravity. I couldn't get anything decent. I have no visibility on what everyone else is doing — what's the best way? How much time does that take? What's the stack?"

The work isn't the prompt. It's the cookbook around it: brand bible, spec cards, asset-type templates, scoring agents, QC loops, change logs. That's what gets installed.

▶ WALT · ON THE LIFT

"We don't want to fire the creative strategist — that's my wife. We want to make her job faster. Smaller clients, smaller budgets. We need leverage. Speed and simplicity. Get the ball rolling rather than building from total scratch."

The system is calibrated for exactly that: a strategist with taste in front, a markdown-file production line behind. Volume goes up without headcount going up.

TL;DR: You don't need a course. You need someone who has already broken every part of this on real client work to install the working version inside Media Made Simple and stay close while you run it.
CHAPTER II · THE BET

Three reasons this works for an agency, not just a brand.

▶ THESIS 01

Volume + diversity is the unlock

One perfect ad doesn't move the needle. 14 unique statics a week, with the right spread of personas, angles, and formats, does. Andromeda needs the variance. The system was built around feeding it.

At Selfmade, the AI statics now outperform what our editors ship. Not by a little.

▶ THESIS 02

You don't need a course. You need an install.

You're already savvy with Claude Code, GitHub, building apps. A course wastes that. The work is in the markdown files, the asset-type templates, the agent loops — and getting them tuned to your way of working without re-litigating every decision.

This engagement skips the teaching and goes straight to building it with you.

▶ THESIS 03

The cookbook compounds across clients

Every Media Made Simple client you onboard gets the same brand bible flow, the same ad-tagging pipeline, the same QC pattern. Each one teaches the system something the next client benefits from.

One install. Then it pays you back every time you sign a new DTC account.

CHAPTER III · THE STACK

What gets installed inside your repo.

Three tranches. Foundation lands first because nothing downstream works without it. Production and Optimization layer on as the system starts shipping.

M1

Foundation

▸ INSTALLS WEEKS 1–2
  • Meta API tagging pipeline — every ad in the account auto-tagged by persona, angle, emotion, asset type, format
  • Adlib gap analysis — over- and under-indexing across your client and their category
  • Historical winner analysis — surfaces what worked, why, and feeds it downstream
  • Brand bible + spec cards — generated from a brand's site in 30 minutes per client
  • The 50-universal-personas framework so analysis is comparable across accounts
M1

M2

Production

▸ INSTALLS WEEKS 3–6
  • Brief-generation pipeline — 10 copy agents iterate each brief until it scores 90+
  • Brief-to-prompt converter — turns approved briefs into the structured prompt format GPT-Image and Nano Banana actually want
  • AI static generation — fal.ai integration, variant generation, folder organization, cost tracking, all from inside Claude Code
  • 15 asset-type markdown files (before/after, social proof, founder story, features/benefits, us-vs-them, etc.)
  • Visual style cards + brand spec cards in the prompt assembly
M2

M3

Optimization

▸ INSTALLS WEEKS 7–12
  • QC agent loop — reviews each generated static against a 50-point checklist, regenerates until it passes
  • Human-feedback markdown updater — when your strategist flags an issue, the relevant markdown file gets updated so the system stops repeating it
  • Brand-voice agent — trained on each client's winning ads, makes every future brief sound like them
  • Change-log workflow — read-write log so your team can flag fixes without stepping on each other's edits
  • Handoff doc + "running it without me" checklist at month 3 close
CHAPTER IV · HOW WE WORK

Sprint in month 1. Calibration in months 2 and 3.

Higher cadence in the build month. Lower cadence after, when the system is running and just needs nudging. This is the shape we discussed on the call — month sprint of weekly calls, then a monthly touch-base with Slack updates as I figure new things out.

CHAPTER V · INVESTMENT

One line item. Three months.

$14,0003 months committed · billed monthly · USD
MONTH 01 · SPRINT
$6,000
MONTH 02 · CALIBRATE
$4,000
MONTH 03 · CALIBRATE
$4,000

Heavier in month 1 because that's where the build happens. Lower after because the system runs itself with light calibration once it's in. 3-month minimum committed up front. No long-term lock-in beyond it.

What happens after month 3?
Optional ongoing collab at $2,500/mo — Slack access, kit drops, and one 60-min call per month. Month-to-month, 30 days notice to cancel. Kit drops and Adlib access end on cancel. You keep every file already on your machines.
FINE PRINT

What's not in the box.

So nobody's surprised in month 2. If any of these become important, we scope them separately.

  • Media buying or ad-account management — your team keeps that
  • Final QC on shipped ads — your humans stay in the loop
  • Copywriting on your behalf — this makes your team faster, it doesn't replace them
  • Work for direct competitors of your active Media Made Simple clients — mutual exclusivity clause in the SOW
CHAPTER VI · KICKOFF

From signed to first AI statics shipped in 30 days.

The sprint pace. Each item below is what's true at the end of that week.

▶ DAY 0

Signed + Stripe.

SOW countersigned, first invoice paid. Slack channel created. Kickoff call scheduled inside 48 hours.

▶ WEEK 01

Access + first brand bible.

Meta API access + Adlib access provisioned. Brand bible and spec cards drafted from your first MMS test client's site. Repo skeleton committed.

▶ WEEK 02

Foundation workflows installed.

Tagging pipeline, gap analysis, and historical winner analysis live in your repo. First gap-analysis dashboard for the MMS test client lands in Slack.

▶ WEEK 03

Production pipeline live.

Brief-generation agents, prompt converter, and fal.ai static generation all running. First test briefs go through the 10-agent scoring loop.

▶ WEEK 04 · SPRINT END

First AI statics shipped to a real client.

14 statics generated, QC'd, and delivered to your first Media Made Simple client. Sprint cadence ends. We move to bi-weekly + Slack for months 2–3.

▶ MONTHS 02–03

Calibration. Optimization tranche. Handoff.

QC agent loop, human-feedback markdown updater, and per-client brand-voice agent installed. Kit drops as I ship new asset types and model integrations. Month 3 closes with a "running it without me" checklist.

▶ READY TO START? ◀

Let's get the
first statics shipped.

If this is a go, I send the SOW + Stripe link inside 24 hours. Kickoff call scheduled inside 48. First AI statics in your client's account inside 30 days.

Will Sartorius
— Will Sartorius · April 2026