How Flo Runs $1,500/Day in Meta Ads Without Opening the Ads Manager
"Now I work with agents for monitoring competitors, creating static ads, publishing them, and analyzing their performance. This has 10x'd my capabilities to iterate on ads."
Flo, Founder of Notis AI
The Problem: $1,500/Day in Ads, Zero Desire to Manage Them
Flo spends $1,500+/day on Meta Ads for Notis AI. Paid ads are his main acquisition channel. But he hates every second he spends in the Ads Manager.
He works with AI agents all day for coding and product. Then he switches to Meta's interface and the contrast is brutal.
"It's horrible when you work a lot with agents, especially for coding, and then you have to use an interface like Meta Ads. You feel like you're entering some kind of time warp where suddenly you're 60 times less productive."
The Meta Ads UI is so fragile that Flo accidentally deleted his best-performing campaign with a single click. The one bringing the most conversions over the last two weeks, gone.
"Every time I breathe on it, everything breaks."
He knows he needs to constantly test new creatives. Meta's algorithm needs fresh ads to optimize. But the process of creating campaigns, configuring ad sets, uploading creatives, and managing budgets across countries is so painful that he just doesn't do it enough.
"You need to feed the beast. You need to add creatives, you need to let the algorithm do its job, and today I'm not doing it."
He spends 80% of his time on product and 20% on ads, even though ads drive his revenue. Not because he doesn't care, but because the tooling makes him dread the work.
Why Other Solutions Don't Work
Flo tried multiple alternatives before AdKit. None of them solved the core problem.
Browser agents (like letting Claude control the Ads Manager directly) get your account banned. Facebook detects bot-like behavior on Business Manager and flags it immediately. This isn't theoretical, it's a known risk that Meta enforces aggressively.
Other MCPs and API wrappers let you read your ad data, but that's only half the job. The moment you try to create a campaign, things break. Meta's campaign structure has strict dependencies: campaigns contain ad sets, ad sets contain ads, each layer has its own targeting, budget, and creative rules. One wrong API call on a live account can waste real money.
"If you've ever tried to do this with the Meta API or any other MCP, you'll quickly realize that they might give you access to your performance data. But the moment you start creating campaigns, everything will break."
AdKit handles this differently. It abstracts the API complexity so your agent doesn't need to understand Meta's campaign hierarchy. And critically, everything goes through draft mode first. Nothing touches your Facebook account until you explicitly approve it.
"Nothing goes to Facebook as long as you don't approve it yourself."
That safety layer is what makes it possible to let AI agents manage real ad spend. Without it, you're one hallucination away from blowing your daily budget on a broken campaign.
What Flo Built
Flo connected AdKit's MCP to his Notis agent stack (alongside Claude Code, Codex, and Cursor). He built a workflow where his AI agents handle the entire ad cycle:
- Find competitor ads in AdKit's ad library
- Save inspiration to his swipe file
- Generate creative variations adapted for Notis
- Create campaigns and ad sets with proper structure
- Review and approve via draft mode before anything goes live
- Analyze performance and identify what's working
- Create new variations based on the results
All of this happens without opening Meta Ads Manager once.
"Now I can use Notis to generate an ad based on the most performing ads from my competitor, and then create the campaign, and then analyze the performance of this campaign, and then create new ads based on the key learnings."
"Nico created something that just works like magic."
The Result
"Now I work with agents for monitoring competitors, creating static ads, publishing them, and analyzing their performance. This has 10x'd my capabilities to iterate on ads."
Flo went from dreading ad management to running it on autopilot. His agents manage $1,500+/day in Meta Ads spend. He focuses on product and strategy while the toolchain handles execution.
The key shift wasn't just automation. It was removing the reason he avoided the work in the first place. When creating a new campaign takes a conversation with an AI agent instead of 45 minutes clicking through Meta's UI, you actually do it.
A hidden benefit of AdKit is that it's built by a solo founder. Flo experienced that firsthand: hands-on support to fine-tune his setup, and a product that's built around how founders actually work.
"He's a solo founder like me and is 100% behind his product. It's really a joy to work with him."
So, if you want to start saving time every week like Flo, try Adkit 😄
Ready to automate your ad workflow?
Try AdKit