My AI employee
Ads, ready for your approval.
Susi has been running in my business since July 17, 2026, and there's a ready-made template for her in Premium.
I give Susi a goal: community, webinar or retargeting. She comes back with a complete campaign: a competitor research pass through the Ad Library, a map built from real audience wording, copy following a fixed four-step structure, images as static HTML plus an AI image layer, and a campaign already set up in the Meta account, forced into paused status by design. I'm the one who flips it live.
47 reports
Reports since launch
3,000–4,000 impressions
Process value before a verdict
Paused
Forced starting state
AI ads employee
A real workday
Numbers from Susi's own course and from 47 reports of her real operation since July 17, 2026.
47 reports
Reports since launch
3,000 to 4,000 impressions
Test window before any verdict
€60/purchase
Kill threshold
Her own course (Lesson 12, "Measuring and Deciding") runs one fixed rule through every campaign: Susi tests for at least three full days or 3,000 to 4,000 impressions before passing a verdict, and decides using fixed if-then rules: kill, scale by 20 percent, or keep running, never by gut feel.
In real operation, she has carried that rule through under real pressure. The closing report from September 1, 2026 shows two Claude Community purchase campaigns that had together spent €705.99 for seven paying new members, €100.86 per purchase, above the hard €60 kill threshold set on August 30, 2026. Susi's verdict: kill, pause, with the reasoning that the best hook isn't dead, it just ran out of its test window. On September 19, 2026 she also extended her own analytics tool, because a change to the Skool links surfaced a new traffic source her script didn't recognize yet: she added it to the source list, left the existing table's column order untouched so old rows stay readable, and ran her 15 automated tests green afterward.
Susi for your business
Your offer needs different ads. Susi's testing logic stays.
Susi does not copy my community campaign. You give her the objective, audience, platform rules, budget boundaries and the metric she should use to evaluate performance.
How I use her
Research, build and evaluate community campaigns
Susi develops ad variations, creates campaigns in a paused state and evaluates them against a defined test window and explicit kill or scale rules.
Adapted to you
Your funnel and your success metric
You replace the offer, audience, conversion goal and thresholds. Susi turns those into suitable research, ad logic and reporting without ever launching a campaign on her own.
What you change
- Offer and audience
- Conversion goal and funnel
- Budget, test window and thresholds
- Approval and reporting
Other possible uses
- Webinar registration
- Lead magnet
- Retargeting
- Local offer
The process stays: Research the market and its language first, test variants with control, always start paused and decide against rules set in advance.
The tools she works with
Susi has her own course in my community and her own analytics tool that she maintains herself:
Understand competitors first
Ad Library research on competitors as the first step of every campaign.
Map audience language
An audience map built from real target-audience wording instead of guesses.
Build ads systematically
Copy written to a fixed four-step structure, images as static HTML plus an AI image layer.
Evaluate results reliably
Her own analytics script, which she extends herself and secures with automated tests, most recently 15 of 15 green on September 19, 2026.
What she can't do, where I approve
She never spends a euro herself. Every campaign the Meta ads connector creates starts forced into paused status by design. Flipping it live is my move alone.
Nothing goes live externally on her own: no emails, no live campaigns, no budget or entity changes without my explicit approval in that specific case. Reading is always allowed.
No invented numbers. If data is missing, she says so instead of estimating.
Attribution runs strictly on Hyros, not on Meta's own attribution; Meta only serves delivery diagnostics (CPM, CTR, hook rate).
Meta's advertising rules are mandatory: no personal attributes in targeting, no unrealistic promises, AI images get labeled.
Learn the process. Or load Susi directly.
Standard
Standard: build it yourself
$5a month
Learn the structure, rules and approval gates, then adapt Susi to your own business.
- Understand the build step by step
- Adapt rules and examples to your own business
Premium
Ready-made templatePremium: ready-made template
$149a month
Susi is a ready-made package in the community.
- Load Susi as a proven employee folder
- Adapt the template to your workflows and approvals
- Community and course in one shared workspace
As of October 29, 2026
Frequently asked questions about Susi
Can AI create Meta Ads for you?
Yes, for the groundwork: Susi researches competitors' Ad Library entries, builds an audience map from real target-audience wording, and writes copy following a fixed four-step structure. She sets up the campaign forced into paused status. Flipping it live and the final call stay with me, she doesn't replace that.
What does an AI agency do differently for Meta Ads than a classic ads agency?
Susi isn't an agency, she's an employee with a fixed tool chain: the same research steps, the same if-then rules for every campaign, documented instead of based on gut feeling. A classic agency sells hours and individual experience, Susi delivers the same process every time, plus reports that justify her verdicts with numbers.
Can an AI set up an entire ad campaign on its own?
Technically yes, with a clear boundary: Susi builds the audience, copy and images and creates the campaign in the Meta account, but every campaign starts forced into paused status. She never releases a single euro herself, budget and live decisions stay with me alone.
What does it cost to run Facebook ads with AI?
The tool itself costs, through my community, from $5 a month for the self-build path or $149 for the ready-made Susi template (as of October 29, 2026). On top comes the actual ad budget at Meta, which I set and approve myself, independent of the tool.
Does AI replace an ads agency completely?
No. Susi handles research, setup and copy, but the decision whether a campaign goes live, and approving the budget, stay with me. She replaces the work beforehand, not the business decision afterward.
Want a campaign that never goes live unchecked?
Susi's template is in the community, with Ad Library research, the four-step copy structure and the paused-by-design rule ready to rebuild.
Join the communityRead on
September 22, 2026 · 13 min read
Buyer Persona with AI from Community Data
Buyer persona with AI from community data: compare buyers with non-buyers, back every claim with a source, know the limits. The method from practice.
Read the article: Buyer Persona with AI from Community DataSeptember 21, 2026 · 14 min read
Skool Community: Free or Paid?
Skool free vs paid: when a free community fits, when a paid one pays off, how money and fees flow on Skool, and why I switched my own community to paid.
Read the article: Skool Community: Free or Paid?September 17, 2026 · 16 min read
What a Meta Ads Agency Costs
What a Meta ads agency costs: five billing models and the incentives they create, which costs sit outside the fee, and how to spot a weak proposal.
Read the article: What a Meta Ads Agency CostsAlso:Creating Meta Ads With AI: A Guide·Set Up a Meta Ad Account and Campaign·Skool Analytics: Reading Posts and Members·Skool API and Automation: What Works·Skool Review from an Operator's Seat·Skool Growth Boost: Bidding Up to 70%·Running Meta Ads to a Skool Community·How to Grow a Skool Community·Skool Onboarding: The First Seven Days·How to Run Ads on Meta: A Guide·Ad Copy With AI: Creatives You Can Test·Audience Analysis for Ads: A Guide·Learn Claude: The Path in Five Stages·Hiring an AI Employee: The Process·Creating an AI Employee: How to Start·AI Employee or Virtual Assistant?·AI and Privacy: What the AI Gets to See·My AI Workforce: Fourteen Employees, One Boss·Hire Your First AI Employee in One Hour·Back to the org chart