# Lesson 5 — Plan → Distribute → Discover

Outcome: a versioned content strategy, channel-specific campaign previews and an evidence-based AEO recheck. Actual demo scheduling additionally requires the isolated canonical Social backend capability; no course request authorizes production publication.

## Before class and five-minute start

Install `--lesson 5` with shared setup. Use separate workspaces and synthetic content. Follow [Plan](../../content-strategy/README.md) to preview/save the four-topic strategy, then [Distribute](../../prompthon-social-campaign-manager/README.md) to create the local two-post plan. No external channel is used. Python needs IANA timezone data (install tzdata on systems without it); Node is used only for the optional synthetic browser-adapter contract harness.

## 25-minute exercise

1. Minutes 0–8: challenge the goal, audience, pillars, searchable/shareable choices and score assumptions. Save revision 1, modify one judgment and save revision 2 under the same strategy id.
2. Minutes 8–16: adapt a core topic into two channel voices. Prepare the campaign, inspect strategy revision, dates, provider IDs and plan hash. Without live isolated-demo capability, stop at prepared and clearly report that no canonical posts/schedules exist. With verified course backend access, separately approve DRAFT then SCHEDULE and read back canonical evidence; production remains forbidden.
3. Minutes 16–25: run [Discover](../../ai-search-visibility/README.md), inspect the five intentionally seeded findings, correct a copy against the [synthetic fact brief](../../content-strategy/examples/synthetic-workshop-brief.md), then recheck the same audit id/revision. Its explicit build deliverable supports the missing answer. Four supported edits resolve four findings; missing evidence attribution stays open without a relevant reviewed source. Compare resolved versus still-open findings; changed query scope must not count as a fix.

Modification: change one provider's copy (new plan hash) or remove a query (scope change, not resolution). Do not invent demand data or promise a search rank/model citation.

## Instructor evidence and reset

Collect strategy revisions/calendar, campaign preview/hash with false canonical-object flag, AEO source hashes/snippets and recheck status. If live demo services are available, collect actual Host receipts, canonical post/schedule/simulation evidence and course-run readback as a separate state. Mock contract results are not deployed-service evidence.

Shared reset clears only scoped course records; it does not delete canonical Social campaigns/posts, source pages, local previews or provider content. Keep those for recovery; use an explicitly authorized owning-app cleanup if necessary. The backend dependency remains issue #221 until independently provisioned and verified.