Three deliverables per founder.
One canvas the operator iterates from.
Audience Onboard is the workflow that turns a signed founder into a shippable LinkedIn rewrite, a 16-week reshare calendar, and a set of per-segment outbound campaigns. The skills draft. The operator decides. The dashboard is the canvas the operator iterates from before anything reaches the client.
Identity. Narrative. Audience.
Every founder onboard produces the same three artifacts. Each one ladders into the next: the voice profile from Identity shapes the captions in Narrative, which shapes the messages in Audience. Three deliverables, one founder voice underneath.
Intake → Build → Export → Revise.
Every deliverable follows the same four-phase rhythm. The first three phases run in sequence to ship the initial draft. The fourth phase — /revise— is the iteration loop that runs whenever the operator surfaces a change from a dashboard review or a client feedback call. There is no "final" revise; the deliverable evolves until the client signs off and execution begins.
Every phase ends at a human checkpoint. The operator approves the strategy before content gets drafted, reviews the content before it ships to the dashboard, and drives every revise pass after that. The skills are good at structure and consistency. The operator is good at judgment. The rhythm exists so neither one does the other's job.
Each deliverable is its own intake → build → export chain.
The four-phase rhythm applies three times — once per deliverable. Each sub-workflow has its own slash-command skills, its own approved strategy doc, its own outputs. The voice profile from Profile Identity is shared downstream so the founder sounds like the founder across the calendar and the campaigns.
Per-segment outbound campaigns the operator ships through Audience
Decides Strategy A/B/C/D/E per segment, then synthesizes the full Stage 1 / Stage 2 / Stage 3 message set per segment — plus a video script when video is in scope. Reads the founder's business profile heavily for differentiation rules and founder-verbatim pain quotes.
/campaign-intakePer-segment strategy decision (A/B/C/D/E) + features + positioning/campaign-buildSynthesize Stage 1 / Stage 2 / Stage 3 + video script per segment/campaign-exportCompile to dashboard JSON + Google-Doc-ready backup
- workspace/campaign-dev/campaign-strategy-approved.md
- outputs/profile-identity/voice-profile.md
- audience-acquisition-strategy.md
- business-profile.md (§4 anti-positioning + §6 founder-verbatim quotes)
- onboard-form.md / call-transcripts/*.md
- reference/campaign-library/playbook.md
- workspace/campaign-dev/campaign-strategy-approved.md
- outputs/campaign-dev/<segment-slug>.md (one per segment)
- outputs/campaign-dev/video-script.md (if video in scope)
- founder-dashboards/prospects/<id>/audience.json
- outputs/dashboard-export/audience-deliverable.md
Operator approves campaign-strategy-review → campaign-strategy-approved before /campaign-build runs. Build's cross-segment voice consistency check fires before any output file is written. Export validates JSON against the dashboard TypeScript schema before writing — a malformed JSON would crash the dashboard route.
One operator observation. Four artifacts updated in sync.
Revise is workflow-level, not sub-workflow-specific. The operator pastes feedback in one of three shapes — free-form observations from a dashboard review, a structured edit list from the team or client, or a transcript from a client review call — and the skill cascades the changes through every affected artifact atomically. The dashboard never goes out of sync with the source markdown; the markdown backup never goes out of sync with the dashboard. The cascade is the work.
Free-form observations, a structured edit list, or a client call transcript file path. Multiple modes can be combined in one pass.
Each change item gets tagged: identity / narrative / audience / preflight. Ambiguous items become clarifying questions, not invented fixes.
Before/after for every affected file. Cascade scope listed up front. Operator can adjust before commit; the cascade is hard to unwind.
Source markdowns updated → outputs regenerated → dashboard JSON recomposed → markdown backup rewritten. Change log entries appended to each affected approved doc.
Stale files everywhere
Operator updates the dashboard inline. Then has to remember to update the source markdown. Then has to regenerate the markdown backup for the client doc. Half the time something gets missed; the next session reads a stale strategy doc and re-derives outdated decisions.
Source-first, cascade always
Source markdowns are authoritative; the dashboard JSON and markdown backup are derivatives. Revise updates the source first, then regenerates the derivatives in one pass. The Change log section in each approved doc captures lightweight history. No -v2, no -current, no drift.
The guardrails every skill enforces.
These five rules are baked into every skill in the workflow — intake, build, export, revise. They're what keeps the deliverables shippable when the workflow is run by a new operator, on a new founder, against a new segment.
No -v2, no -current, no parallel drafts. The approved doc is the live doc; revise updates in place. Change log sections capture lightweight history when the operator wants it.
psychological-profile.md exists to help Future Media convert leads into clients. Reading it during campaign-dev pollutes the outputs with sales-conversion signals. Verbatim founder quotes come from onboard-form, intake-approved, and business-profile instead.
Final client deliverables ship one best version per section. The operator decides; the client executes. The A/B variants on connection requests are the documented exception per playbook §4.b — an operator-driven test, not a buffet of hedge options.
Every number, named past employer, named client, status signal, and asset link in the copy must trace to a source extraction. If the approved doc says "20+ years" but the intake says 18, use 18 and flag. Fabrication kills trust on the first send and is unrecoverable.
The universal calibration rule: when a segment is broad, cold, and multiplier-tier (influencers / analysts / editors — not direct buyers) AND the founder has a strong media asset (podcast OR newsletter), default to Strategy B over C. Direct buyer segments still default to C per playbook §5.a.
The workflow is the leverage. The operator is the judgment.
None of this is autopilot. The skills produce structure — consistent file shapes, voice-checked copy, validated dashboard JSON. The operator produces judgment — which strategy fits a segment, when a client's feedback warrants a strategy flip, whether a Stage 2 message reads like the founder or like an agent.
The workflow exists so the operator's judgment compounds across every founder — not so the operator can be replaced by it.