An AI Performance Creative Director that runs the entire creative testing machine. Audits my account every Monday. Writes up to 15 production briefs every Tuesday across 4 tracks (AI, UGC, new shoots, graphics). Pulls competitor intel from the Meta Ad Library and signals from the 8 top DTC performance marketers. The work of a $10K to $20K/month creative agency, running on my own account.
Why I Built This
Most brands have one creative strategist. They write one brief a week. They get sick once a quarter. They quit within 18 months. The brand starves between hires.
Meanwhile, the actual work of creative strategy is mostly the kind of thing AI is built for. Audit the account. Compare every winner to your benchmarks. Spot the ones that are fatiguing. Mine the patterns. Write the next brief with the hypothesis attached. Track every winner and every failure in a database so you never re-test what already lost. Watch the competitors. Read what the top 8 DTC performance marketers are publishing. Synthesize.
No human can do that volume on a schedule. A real creative strategist costs $10,000 to $20,000 a month and still only ships one or two briefs a week. The math doesn’t work for any brand under eight figures.
So I built the system I wanted. An AI Creative Director that never sleeps, never quits, and runs the entire creative machine on a schedule. Monday audit. Tuesday brief drop. 10 to 15 production-ready briefs every single week. Every brief comes with the hypothesis attached. Every brief references your proven winners. Volume scales to your spend. The account gets smarter every week.
What This Took to Build
It took me weeks of building
and thousands of dollars in API tokens
to make this happen.
Not a template. Not a prompt wrapper. The actual production agent I run on my own Meta accounts every week, built so any brand can drop it onto theirs and stop bottlenecking on creative.
Monday
Full weekly creative audit
Pulls every active creative in your account. Scores each one against your own benchmarks for hook rate, CTR, hold rate, CPR, and ROAS. Labels every ad as SCALING, HOLDING, FATIGUING, or KILL. Hands you the list in plain language. The work an agency would do every Monday before they bill you for it.
Tuesday
10 to 15 production-ready briefs
Every Tuesday, a full brief package across 4 tracks: AI automation, real UGC, new content shoots, and graphic design. Every brief has a stated hypothesis. Every brief references your proven winners. Every brief is specific enough that a creator, designer, or shoot team can execute without follow-up questions.
Winners
Winner exploitation, before new concepts
When a creative wins, the agent doesn’t run to the next shiny idea. It milks the winner first. 5 to 10 variations on the same concept: new hooks on the same body, new formats on the same script, new creators on the same template. You extract every dollar of value from the proven concept before introducing untested ones.
Memory
The Winner Database
Every proven concept and every documented failure, logged forever. Three months in, the agent has institutional memory of your account that no human strategist could match. You never re-test a concept that already lost. You never forget why a winner worked.
Benchmarks
24 months of YOUR data, turned into thresholds
At install, the agent pulls 24 months of your account’s creative performance and calculates real percentile thresholds (P25, P50, P75, P90) for every key metric. Your kill threshold and your winner threshold are derived from your own account, not industry averages. Generic Meta benchmarks lie. Your account does not.
Vision
GPT-4o vision SEES your creatives
During install, the agent uses GPT-4o vision to analyze your top 100 and bottom 50 creatives. Hook type, product visibility, creator type, setting, production quality, visual patterns. Then it writes the rules: what your winners actually have in common, what your losers share. Pattern recognition no human strategist could do at this depth.
Competitors
Weekly Meta Ad Library scrape
Every Monday, the agent pulls every active ad from the competitors you watch. What formats they’re running, what hooks they’re testing, what offers are live, which ads have been running for weeks (the proven winners). You get a digest of what they’re doing and where the white space is for you.
Market
The 8 best DTC marketers, synthesized weekly
Every week, the agent scrapes new posts from Savannah Sanchez, Barry Hott, Cody Plofker, Ash Melwani, Taylor Holiday, Nik Sharma, Nick Shackelford, and Satish Prajapati. Plus three DTC subreddits and four creative-strategy blogs. Summarizes the hook formulas, testing frameworks, and platform changes. So you’re briefing with what’s working in the market right now, not last year’s playbook.
Every Tuesday
Up to 15 production briefs.
The same package agencies charge
$10K to $20K/month to deliver.
Across 4 tracks: AI automation, real UGC creators, new content shoots, graphic design. Every brief comes with a hypothesis. Every brief references the proven winners from your own account. Volume scales with your Meta spend, so a $200K/month brand gets 5 to 7 briefs a week and a $600K+ brand gets 11 to 15. Production-ready. No follow-up questions. Hand it to your designer, your UGC creator, or your editor and they can execute.
Most brands ship 4 to 8 new creatives a month. This is 4 to 8 a week. The account gets smarter every single Tuesday.
Get the Agent
The complete agent. Context files, the brief templates for all 4 tracks, the weekly audit workflow, the winner exploitation framework, and the full intel scraping system (Meta Ad Library + Twitter + Reddit + DTC blogs). Drop it into Claude Code and your first brief package lands this week.
Get the Creative Director Agent →One-time purchase. Runs in your own Claude Code. You own it.
Go All In
If this guide helped, but you’re looking to go deeper, I got you!! My 30-Day Challenge takes you from saving AI tips you never use to actually building with AI, step-by-step.
I show you exactly how I automated two e-commerce brands, my social media, and most of my personal life, then hand you the agents, workflows & systems to do the same. I’m teaching you every single thing I know with one lesson and one build a day.
Join the Challenge →Doors close June 21. We start June 22.
© 2026 Mariah Brunner. All rights reserved.