Diagnose, then campaigns
What it does: Turns customer language (and your URL) into 6 Ps positioning, campaign angles, a focused acquisition plan, weekly content, and an optional launch checklist. After onboarding, Mark sends you to Diagnose first.
Try this first: Paste CRM, tickets, or notes in Diagnose, pick an angle, then generate one strategy. Open that campaign to execute; use Content and Calendar filters when browsing drafts.
Pro tip: Constrain your audience clearly: buyer role, company stage, and top pain. Specific inputs produce specific strategy.
Common mistake: Opening a blank campaign wizard before you have buyer language. You get generic output.
Content, Preview, and ship plan
What it does: Drafts channel-ready posts from strategy, shows a founder-facing Preview (hook, body, CTA — not raw JSON), routes approved pieces to that campaign’s ship plan, and supports scheduled or API posting. Preview also flags extractability for GEO pages.
Try this first: From a campaign home, generate posts, read them in Preview, approve, then publish from that plan’s Ship plan. Home can show all campaigns when triaging.
Pro tip: Keep tone constraints short and explicit. Add one real proof point in each draft before posting.
Common mistake: Opening Ship plan from Home for all campaigns when you meant to ship one plan in order.
Inbox and Home tracking
What it does: Home ranks posting, inbox threads, AI visibility, competitors, outcomes, and experiments. Inbox surfaces Reddit, X, and Indie Hackers threads with WARM reply drafts and freshness windows. On iOS use the Engage tab (Inbox, Diagnose, Memory). Posts due and failed posts count the full library.
Try this first: Check the Tracking card each morning; reply to the highest-priority conversation before the window closes.
Pro tip: Use objection language from replies to refine next week prompts and landing copy.
Common mistake: Treating conversations like a link-drop workflow. Promotion-first replies get ignored.
Growth OS
What it does: Marketing memory for the workspace: customer truth, founder voice, winning hooks, outbound triggers, agent specs, and transcript→assets so each generation cycle gets sharper.
Try this first: After your first Diagnose and a week of outcomes, open Growth OS and hit Refresh memory.
Pro tip: Promote winning experiment hooks into Growth OS instead of rewriting the same angle from scratch.
Common mistake: Treating Growth OS as a report you read once. It only compounds if you refresh after real customer data.
AI visibility (GEO)
What it does: Measures which buyer questions name you across AI engines, shows citation rank, classifies silent vs referenced mentions, lists lost questions, and builds a 7–15 day citation sprint.
Try this first: Run your first measurement from AI Visibility, pick the worst lost question, ship an extractable page, then re-measure.
Pro tip: Ship comparison and checklist content when you lose a question — follow the GEO content checklist and Preview extractability checks.
Common mistake: Measuring once and never shipping the lost-question page.
Landing Audit and analytics
What it does: Landing Audit scores a URL and lists prioritized fixes. Analytics tracks outcomes, runs A/B angle tests per campaign, derives learning notes, and feeds insights into Growth OS.
Try this first: Audit your homepage once, log outcomes weekly, and launch one experiment with a minimum sample per variant.
Pro tip: Judge posts by intent and outcomes, not impressions. Save only lessons that change future decisions.
Common mistake: Looking at analytics without converting insights into new prompt constraints and channel rules.
Integrations, iOS, and execution
What it does: Connects Bluesky, email, analytics, and posting so publishing stays in one workspace. iOS TestFlight loads the same live Mark — last workspace restores on open. Archive a campaign to hide it; delete only after typing the name. Workspaces hide for 30 days before purge.
Try this first: Connect one destination first and run a small batch from the ship plan to confirm format and naming.
Pro tip: Standardize naming for campaigns and assets so weekly reviews are readable.
Common mistake: Automating distribution before message quality is stable. You scale noise, not results.