Frequently asked questions
What's the real difference between the two modes?
Autonomous = the agent acts on its own within guardrails you set (max spend changes, what can ship without review, etc.). Semi-autonomous = same thinking, but every move queues for you to approve before execution. Same intelligence, different level of human-in-the-loop. You can mix them — autonomous on creative, semi on budget.
I want to stay in control of my ads.
Use semi-autonomous mode. Every proposed change shows up with reasoning, a before/after diff, and projected impact. You approve, reject, or edit. Plus every action — in either mode — is logged and reversible in one click.
I already have an ads agency.
The agent doesn't replace strategy. It replaces the execution layer the agency bills you for — monitoring, adjusting, reporting. Keep the strategy. Cut the overhead.
How is this different from Meta's own automation?
Meta's automation optimises for Meta's outcomes. The agent optimises for yours — your margins, your bestsellers, your growth targets. It also looks beyond the API: competitor Ad Library, trending keywords on Reels, your Shopify data — and factors them in.
What if it does something I don't like?
Every change is reversible from the decision log in one click. You can tighten the guardrail that allowed it, and the agent learns the new boundary immediately. In semi-autonomous mode, rejections also train it.
How fast can I get it running?
Connect Meta, import your store, set guardrails — most stores are live within 30 minutes. The agent runs in shadow-mode for the first 7 days: tells you what it would have done, but does not act yet. You flip it on when you trust it.