How Predicai is Designed

How Predicai
is built.

Transparency about the systems, decisions, and design choices behind Predicai. No mystery. No black box.

The Curriculum

Built on frameworks, not tools.

Every lesson is built around reusable thinking frameworks — not specific AI tools. Tools change. Frameworks survive. The Intent Protocol, the 7 Gates, and the Agentic Blueprint are designed to hold up across model generations.

The five-level capability system is universal — the levels and standard do not change. Role Architecture defines what Gold looks like inside each function. Role Packs for Sales, Chief of Staff, Operations, RevOps, and HR map the same curriculum to role-specific work without redesigning the system. Levels 1–3 (Builder Track) build the workflow. Levels 4–5 and the Capstone (Advanced Track) harden it under real-world pressure. See how it works →

Cai — the Academy co-architect — is built on the Claude API with the full Predicai curriculum as its system context. It enforces the standard on every submission before a human ever sees the work.

The review flow: a student submits an artifact inside their Notion workspace. Cai checks it against the standard and surfaces structured feedback. A human Predicai architect reviews for judgment, nuance, and transferability. The unlock happens when the work holds up — not before.

CurriculumFramework-firstTool-agnostic
CaiClaude API + custom contextCurriculum embedded
WorkspaceNotionArtifacts + dossier
ReviewHuman architectsJudgment, nuance, transferability
CertificationCapstone System DossierDemonstrates working capability
Signal

Written by humans.
Tested in the real world.

Signal is not AI-generated content about AI. It is human observation filtered through the Predicai framework and structured for action.

The filter is simple: does this shift how work actually gets done, and does it hold up past the news cycle? If it is interesting but not operational, it stays out. Each issue includes what changed, why it matters at the system level, and one concrete move. That structure is deliberate — Signal is designed to feed capability, not just awareness.

Signal is released when there is something genuinely worth your time. Never daily. Never to fill a schedule. The editorial discipline is part of the product. Inside the Predicai system, Signal is the live evidence layer — the ongoing record of what is actually changing and what that means for people building with AI.

FormatIntelligence briefingNot a newsletter
Source filterReal-world testingWhat held up in use
CadenceDeliberate releaseQuality over frequency
RoleFeeds capabilityEvidence layer for Academy
CostFree alwaysNo paid tier
What is human vs automated

The system keeps judgment
where it belongs.

Automation surfaces. Humans decide.

AI guidance
Cai
Standards enforcement on first pass. Structured feedback inside the student workspace. Flags submissions that do not meet the 7 Gates before human review begins.
Human review
Predicai architects
Review judgment, nuance, and real-world transferability. The unlock decision is always human. Cai surfaces; architects verify. Review is designed for active evaluation, not passive approval.
Workspace
Notion + artifacts
Persistent context across all five levels. Artifacts compound into the System Dossier. The workspace is the record of capability, not just completion.
Quality layer
The Standard
Shared evaluation criteria across all levels, roles, and use cases. Published openly. The same standard governs Cai guidance and human review.
The common layer

Built to make capability
durable, not trendy.

Across Academy and Signal, the same principle holds: human judgment, strong systems, and explicit review. The goal is not to make AI feel impressive. The goal is to make people more capable and harder to fool — before the next tool cycle starts again.

Explore Academy → Read the Vision →