Filter by track or search by topic (e.g., MCP, Brand Voice, SEO, Carousel, Cohort, Autonomy). Every lesson is available in full HTML in both English and Polish.
LESSON 01TRACK 01Foundations9 min read
How to structure brand strategy as an evolving, versioned operating contract (v1 → measured cohort → v2) that actively governs every content format.
- A strategy stored in a slide deck dies the moment daily publishing begins because writers and AI prompts cannot query it.
- Versioned strategies (v1, v2) treat positioning, content pillars, and channel cadences like software releases tied to measurable 30-day cohorts.
LESSON 02TRACK 01Foundations10 min read
Why adjectives like "friendly and professional" produce generic AI slop—and how to build enforceable voice specifications.
- Adjectives ("modern, authentic, bold") mean nothing to an LLM; structural rules and negative boundaries ("never use hype superlatives") create real voice consistency.
- A complete Brand Voice profile combines Cadence Rules, Domain Vocabulary, Proof Standards, and Forbidden Claims.
LESSON 03TRACK 01Practitioner11 min read
How to eliminate AI hallucinations and generic content by separating raw sources from editorial interpretation (Source → Decision → Draft).
- Never ask an LLM to simultaneously invent facts and write in your brand style; separate source gathering from drafting.
- Curated RSS feeds, official domain allowlists, and existing workspace articles create a high-signal research pool.
LESSON 04TRACK 02Practitioner10 min read
How to turn one researched narrative into a canonical website article, a 5-slide carousel, and native posts for Instagram, Facebook, X, and LinkedIn.
- Publishing 20 disconnected ideas a month creates noise; publishing 4 deeply researched stories adapted across 5 formats builds authority.
- Every channel requires native framing (hook, length, visual ratio, link placement) while sharing 100% of the underlying factual context.
LESSON 05TRACK 02Practitioner9 min read
How to engineer 5-slide educational and product carousels that earn swipes, saves, and qualified traffic.
- A carousel is a sequential micro-article: every slide must resolve one question while opening the loop for the next swipe.
- The canonical 5-slide arc (Hook → Constraint/Problem → Mechanism → Proof in Use → Action/Summary) prevents mid-carousel drop-off.
LESSON 06TRACK 02Architecture11 min read
Why "queued" is not "published"—and how to run resilient multi-target publishing operations without duplicate posts or silent failures.
- A post in a queue is an intention, not a result; external platform APIs fail, rate-limit, or reject media asynchronously.
- Multi-target publications must track state independently per channel so retrying a failed target never duplicates a succeeded one.
LESSON 07TRACK 03Architecture11 min read
How to architect bilingual content hubs, technical SEO metadata, and CMS delivery contracts so social reach builds durable search equity.
- Social media posts decay in 48 hours; your website articles compound in search and AI retrieval for years.
- Bilingual content hubs require strict URL symmetry (`/` and `/pl/`), reciprocal `hreflang` links, and rich `TechArticle` / `WebSite` JSON-LD graphs.
LESSON 08TRACK 03Practitioner10 min read
How to deploy a brand-grounded website chatbot, capture high-intent leads, unify social comments, and mine recurring questions (FAQ Mining).
- A generic website chatbot without scoped brand knowledge frustrates visitors; a grounded assistant answers from curated product docs and workspace articles.
- Audience questions in website chat and social comments are the highest-signal research source for your next article or carousel.
LESSON 09TRACK 03Practitioner10 min read
How to evaluate 7/30/90-day content cohorts across social and web—and turn measured evidence into a reviewable Strategy v2 draft.
- Single-post viral spikes are noisy; 7, 30, and 90-day cohort windows reveal which content pillars and formats consistently compound.
- Always evaluate social engagement alongside website referral traffic and popular pages to see what actually drives deep interest.
LESSON 10TRACK 04Architecture12 min read
How to connect AI agents (Claude Code, Codex, Antigravity) to your workspace via scoped MCP tools instead of copy-pasting into chat windows.
- Copy-pasting strategy docs and analytics CSVs into a browser chat window is brittle, manual, and immediately goes stale.
- Model Context Protocol (MCP) gives AI agents typed, permission-scoped tools to read workspace context, inspect cohorts, and create structured drafts.
LESSON 11TRACK 04Architecture11 min read
Why AI agents should never have a single-step "publish now" tool—and how two-phase confirmation gates prevent accidental live mutations.
- LLMs are probabilistic; external API mutations (publishing a post, replying to a customer, activating a strategy) must be deterministic.
- Splitting every mutation into `prepare_*` (dry-run validation + preview token) and `confirm_*` (token execution) eliminates accidental side effects.
LESSON 12TRACK 04Architecture10 min read
How to configure workspace governance, role-based access, and three graduations of AI autonomy (`draft_only`, `schedule_with_confirmation`, `autonomous_schedule`).
- Autonomy is not a binary on/off switch; teams need graduated modes (`draft_only`, `schedule_with_confirmation`, `autonomous_schedule`) as trust matures.
- Hard operational guardrails—allowed channels, maximum cadence, quiet hours (`22:00–07:00`), and monthly credit budgets—prevent runaway automation.