Searchers type the scary question first: will this thing spend while I sleep? Here is the straight answer.
Automated vs gated
- Automated (read/compute) - syncs, scores, briefs, detectors, caches, drafts
- Gated (write/spend) - campaign publish, platform mutations, bulk fixes that cost money, feed writes where flagged
Why default-off writers exist
Multi-tenant scale plus money-moving APIs is how you get incidents. Flags and pending review are how you sleep.
Related: campaign review, flags list.
How to evaluate any AI ads vendor
- Ask what is draft vs live by default.
- Ask which writers are off until enabled.
- Ask how undo/revert works.
- Ask what happens when inventory is thin.
If answers are vague, the risk is yours.
Questions people actually ask
Does Ralph publish campaigns without me?
Material campaigns are designed to land in pending review until you approve. That is core product philosophy, not fine print.
Are any writers on by default?
Many money-moving writers default off (Google writer, exclusion writers, product set writer, bulk fix, GMC mastery, etc.) until enabled for an environment/tenant.
What about intelligence and previews?
Reading, scoring, drafting, and preview can run without publishing. Publishing and live mutation are the gated step.
Is Concierge allowed to buy ads?
No. Concierge is storefront assistance, not the ads writer.
Product rule: draft and prioritise freely; ship spend only with human approval. Approval docs · Trust.