cida-learn messaging
Copy and narrative structure for the cida-learn home page (https://cida-learn.chrislock.tech/). Designed 2026-08-21. No visual redesign implied; existing demo components are reused and re-captioned. Implementation lives in apps/cida-learn.
Audience contract
One page, two audiences.
- The executive buyer — has never heard of CIDA, Sparqdown, or Markdown. Feels organisational pains: meetings that relitigate settled questions ("which version is valid?", nomenclature fights), documentation produced then forgotten and stale, paying subcontractors to translate between company languages, R&D specialists discovering standards at launch and triggering rework. The hero must hook them in their own vocabulary — agreements, decisions, meetings, rework — never "documentation" as the frame — and route them to the pitch film.
- The practitioner — will actually learn Sparqdown and use the sandbox. Takes over below the fold; the mechanism layer (living documents, four behaviours) speaks to them.
Conversion goals: executive → watch the 3-minute pitch and remember the name. Practitioner → Start learning / Open the sandbox.
Messaging platform
- Umbrella concept (executive layer): what the organisation has agreed — terms, plans, standards, architecture, requirements. Every buyer pain is a failure of agreements: they can't be found, can't be trusted, don't reach the people they govern, expire silently.
- Enemy: the dead document — approved, filed, forgotten, silently wrong. And its cost: the meeting that re-agrees on what was already agreed.
- The genuinely-new claim: wikis and file shares store agreements; they cannot verify them. Mechanical verification plus machine-readability is what technology can finally do.
- Mechanism layer (practitioner): documents that are alive. Four behaviours, each mapping one-to-one to an executive outcome:
- Decisions stay decided ← defined once, cited everywhere (connected corpus).
- Drift surfaces immediately ← claims trace to sources and re-verify.
- Silos can't stay invisible ← documents know each other; rules reach the work structurally.
- AI gets ground truth ← the same source is precise machine-readable knowledge.
- AI position: the intelligence is real; the memory is fake. AI's power is reasoning, not latent knowledge — CIDA is the harness that puts that intelligence to work on explicit, testable knowledge. Neither AI-hype nor AI-bashing.
- Voice: each act opens with an arresting one-liner built on the same engine — an admission every reader knows is true, then a flip. Cadence mixes forms deliberately: question → question-flip → statement (three questions in a row reads as a quiz).
- Name reveal comes last. The CIDA acronym expansion is a payoff for readers who already want the thing, not a headline. Product names (CIDA, Sparqdown) never appear before the visitor has a reason to care.
Page narrative
Hero
Eyebrow (small):
Terms. Versions. Plans. Standards. Decided — then argued about for years.
Headline:
How many meetings does it take to agree on what was already agreed?
Subhead (hinge + inoculation):
What if there were always one current answer — and it defended itself? CIDA makes your organisation's agreements a living, machine-verified source of reality. Not another place to store files. Not a chatbot guessing at PDFs.
Outcome row (four checks):
- Decisions stay decided. One current answer to "which version is right?" — and everyone can point to it.
- Drift surfaces immediately. When reality moves away from the plan, you're warned that day — not the week of launch.
- Silos can't stay invisible. Standards and requirements reach the teams they govern — structurally, not by luck.
- AI gets ground truth. Inference runs on what your company actually agreed — not on guesswork.
CTA row: Watch the pitch · 3 min (primary, executive path) · Start learning · Open the sandbox (secondary, practitioner path).
Choreography (presentation, not copy — the copy block above is unchanged): the hero plays in three beats, not one screen. Beat 1 is the eyebrow and the D1 question alone, centered, with a scroll cue — the visitor's first read is the question and nothing else. One damped scroll past the cue reveals beat 2: the kicker, the hinge + inoculation, the four-outcome row, the CTA row, and the badge strip, laid out beside the constellation/pitch column. Beat 3 is deferred all the way to "Why it's called CIDA": the animated C-I-D-A lockup does not appear in the hero at all — it relocates there as the name-reveal payoff, consistent with the messaging platform's "name reveal comes last" rule.
Notes:
- "Watch the pitch" is promoted from the constellation column into the hero CTA row as the primary action; it stays playable in the right column too.
- The large C-I-D-A typographic treatment is removed from the hero entirely — not even as background/side art — and relocated wholesale to the name-reveal section as its payoff; see the choreography paragraph above.
- The badge row (v0 preview · open spec · local-first · self-describing corpus) stays — practitioner-facing trust markers.
- Constellation animation and chapter chips stay in the right column, framing the pitch film.
Act II — the writer's act (guidance and blind spots)
Header:
Nobody reads the standards folder. What if the standards read your document?
Support copy:
In CIDA, a document knows what it is — a proposal, a spec, a standard — and what your organisation requires of it. The rules come to the writer: required parts, agreed terms, the blind spot you didn't know you had. A specialist can't be unaware of a requirement that is structurally linked to their work — and neither can their AI assistant.
Reuses three existing demo cards from "What working in CIDA feels like", captions kept or lightly executive-framed:
- AI that knows what you're writing — "Since you're writing a project proposal, you must include a project-timeline…" — the flagship demo for this act: the standard literally reading the document.
- Documents that know they're stale — "project-atlas-proposal references cost-baseline v2 — v4 was published last week." — a blind spot covered before it costs anything.
- Claims that re-verify against their sources — "A claim is not consistent with its source." — the watching doesn't stop at publish.
Optional section banner above the cards: Most documents die the day they're approved. These stay alive.
Act III — the machine's act (AI harness)
Header:
AI is a brilliant reasoner with a made-up memory.
Support copy:
Its real power is thinking, not knowing — left on its own, it fills the gaps with statistical guesswork. (Not so unlike people in a meeting.) CIDA is the harness that puts that intelligence to work on your organisation's explicit knowledge — every term defined, every claim testable, every answer traceable to something your company actually agreed. Let AI think. Stop letting it guess.
Reuses two existing demos:
- One document, three faces — what you write / what readers see / what machines get — the mechanism: one source, human-fluent and machine-fluent.
- Knowledge your documents imply — the
reportsToinference demo — reasoning running on explicit knowledge; the harness in action.
Practitioner runway — "How it holds together"
The existing section (Document / Registry / Constellation / Verification plus the verified/diverged/indeterminate legend) stays as-is, now positioned after the acts: the reader arriving here already wants the mechanism.
Name reveal — "Why it's called CIDA"
The existing "Coherence and Integrity" section, recast as the payoff:
Coherence. One source of truth. Values, terms, and definitions live in exactly one place and flow into every page, table, and query that cites them. Documents form a corpus, not a pile — change a fact where it lives and everything that depends on it follows.
Integrity. Claims cite their sources in machine-checkable form. Verification is mechanical, not editorial: every claim is verified, diverged, or indeterminate — and divergence is a warning you see, never a silence you discover too late.
Closing line: Coherence &Integrity — Documentation Architecture. Sparqdown is the language it's written in. Now the names mean something.
Proof and close
- "This page practices what it preaches" stays as-is (prose and knowledge graph from one source; canonical RDF link).
- "Ready when you are" stays; CTA order mirrors the hero (pitch link included for the executive who scrolled).
Copy inventory — canonical one-liners
For reuse in decks, social cards, and the pitch film:
- D1 (hero): How many meetings does it take to agree on what was already agreed?
- Kicker (beat 2 lead): A documentation platform where documents are alive.
- Hinge: What if there were always one current answer — and it defended itself?
- Inoculation: Not another place to store files. Not a chatbot guessing at PDFs.
- E1 (Act II): Nobody reads the standards folder. What if the standards read your document?
- Dead/alive banner: Most documents die the day they're approved. These stay alive.
- F1 (Act III): AI is a brilliant reasoner with a made-up memory.
- F2 (Act III close): Let AI think. Stop letting it guess.
- Four outcomes: Decisions stay decided · Drift surfaces immediately · Silos can't stay invisible · AI gets ground truth.
Deltas vs. the current page — implementation checklist
- Hero headline: acronym expansion → D1 question. The big C-I-D-A lockup is removed from the hero entirely (no visual-signature treatment) and relocated wholesale to "Why it's called CIDA" as the name-reveal payoff.
- Hero eyebrow: "Documentation drifts. Claims go stale. Nothing tells you." → "Terms. Versions. Plans. Standards. Decided — then argued about for years."
- Hero subhead: "CIDA is the architecture. Sparqdown is its language…" → hinge + inoculation sentence (above).
- New four-outcome row in the hero.
- CTA row: add Watch the pitch · 3 min as primary; Start learning / sandbox secondary.
- "What working in CIDA feels like" → split into Act II (three cards) and Act III (two demos) under the new one-liner headers; "One document, three faces" moves into Act III.
- "How it holds together" moves below the acts.
- "Coherence and Integrity" → "Why it's called CIDA" with the closing name-reveal line; acronym expansion lands here.
- No changes to: constellation animation, chapter chips, badges, practices-what-it-preaches, sandbox/learn destinations.
Out of scope
- Pitch film script alignment with the new one-liners (worth doing later — the film's five chapters should echo D1/E1/F1).
- A dedicated "for leaders" page, pricing/contact machinery, and any visual redesign.
- Translation-cost and subcontractor pains: real buyer pains, but too specific for the hero; candidates for a future vignette strip or the pitch film.