Gap-driven, not calendar-driven
The plan is built from your monitored prompts and the questions AI gets — not from a generic posting schedule. Fewer, better-aimed pieces; nothing invented.
Platform
AI visibility monitoring Competitors Brand AEO Score Content Studio Campaigns Search & AI demand Site & page audits Content Gaps Agent ControlResources
Learn
Knowledge center Research Data News & updatesDevelopers
Developer docs Developers & API OpenAPI spec llms.txt Free ToolsPricingSign UpSign InPlatform · Campaigns
A campaign is a time-boxed plan of content for one surface AI already cites. Pick the surface and a cadence, and AEO Spotlight lays out the pieces — each answering a question you actually monitor, in the pattern that works on that surface. You draft, fact-check and publish from your own account; the platform never posts for you, and no outcome is promised.
Real AI queries · no credit card · 3 free checks a day

The plan is built from your monitored prompts and the questions AI gets — not from a generic posting schedule. Fewer, better-aimed pieces; nothing invented.
Each surface has a playbook — voice, structure, link and disclosure conventions, and what AI extracts there. LinkedIn is not Reddit, and the drafts reflect that.
Pick the structure known to work where you publish — X’s “Thread — post + reply”, LinkedIn’s “Insight post (50–299 words)”, Reddit’s “Practitioner answer”. Formats live in an editable catalog, so they can change as the platforms do.
When a topic lands, change the plan in plain language — “Expand on last week’s topic”, “Focus the rest on pricing”, “add 3 posts about onboarding”, or “pause”. You see exactly what will change before it changes, and posted items are never rewritten.
The default is a steady four-week plan. A shorter plan is a burst; you choose it explicitly. Steady cadence is what tends to hold.
A campaign can also schedule refreshes of your decayed, uncited pages — because losing freshness is half of losing citations.
The platform drafts; you publish. There is no post, schedule or seed action, and no platform write of any kind.
Mark an item posted (paste the live URL) and the citation monitor watches whether an AI answer cites it — labelled with the surface’s honest measurability.
Only surfaces you can actually publish to are offered. Wikipedia and Wikidata are earned, not authored, so they are never draftable.
Each surface offers named formats with the mechanic behind each. A thread format schedules the post and its self-reply; a single-post format schedules one piece.
See the whole schedule before you commit — day, item type, the proven pattern, and the question each piece answers.
Drafting runs through Content Studio: grounded in your sources, fact-checked before publishing, then humanised for the surface.
Post from your own account, then paste the live URL so the monitor can watch it.
We can’t promise that, and no tool can. Outcomes depend on third-party AI platforms we do not control. A campaign is a dated plan of actions and a way to watch whether the cited-source picture moves — it is not a guarantee of a mention, citation or ranking.
Never. Campaigns are create-only: the platform drafts, you publish from your own account. This respects platform rules (and several large communities ban automated posting) and keeps every claim verifiable.
LinkedIn, Reddit, YouTube, Quora, review sites (G2/Capterra/Trustpilot), publications, X and other public forums. Wikipedia and Wikidata are earned-only and are listed as goals, not targets.
A format is the named structure for a post on that surface — X’s thread (post the point, then the link in a self-reply), or LinkedIn’s mid-length insight post, for example. Formats are kept in an editable catalog so they can be updated as platform behaviour changes. They are guidance, not a guarantee.
From the visibility gaps you already have: the questions come from your monitored prompts, and the refresh items come from your published pages that have gone stale without being cited. The cadence is a sustainable default you can change.
Mark items posted and the citation monitor watches for AI answers that cite your URLs. The honest scoreboard is share of voice within your own monitored prompt set, per engine, over the campaign window — correlation, never per-post causation.