The Space Between Decisions
Ma.ai maps your organisational knowledge — people, companies, contracts, decisions, risks — into a living graph. Every high-stakes decision is examined through four stages of structured reasoning.
Powered by The Crusible. Built for organisations that cannot afford unexamined assumptions.
Already have access? Sign in
Ontology Knowledge Graph
Maps people, companies, contracts, decisions, risks, and events into a typed relationship graph. Grows organically from every decision run.
Failure Library
Logs post-decision outcomes: outcome failure, process failure, assumption failure. Append-only. Each entry links to the original decision and the assumptions the Crusible surfaced.
Context Declaration
Assembles declared context from the graph and failure library. You confirm, modify, or remove each item before the Crusible call proceeds. Mandatory. Auditable.
The Crusible API
Sovereign four-stage reasoning pipeline. Maps, eliminates, pressure-tests, and reveals. Returns a Revelation. Its internal process is unchanged by Ma.ai.
Two outputs. One run.
The boundary between what is proven and what is inferred is explicit, auditable, and the product.
Correlation Documentation Record
Co-occurrence analysis. Zero causal claims. Structured JSON with hardcoded limitation disclosure. For compliance, due diligence, and legal defence.
Operational Narrative
Plain-language summary inferred from correlation. Explicitly labelled as inference. For board presentations, fiduciary communication, and executive briefings.
The old world left no record.
Ma.ai closes that gap.
Unexamined decisions
Examined. Documented. Defensible.
Every high-stakes decision.
One structured process.
Third-party vendor onboarding. The most common source of enterprise breaches.
Before any vendor is approved, Ma.ai surfaces every assumption your team is making — contractual exposure, integration risk, IR maturity, tool sprawl. The Crusible pressure-tests the decision before it's made.
- Contractual liability exposure examined
- IR maturity gaps surfaced
- Integration failure modes mapped
- Correlation record exportable to auditors
Critical assumption exposed: Your IR runbook assumes 4-hour containment. Vendor SLA is 8-hour response. This gap is undocumented and creates material contractual exposure under three client agreements.
Unexamined risk: Tool sprawl across 7 existing EDR-adjacent products creates integration failure probability your team has not quantified...
Built on The Crusible.
Reasoning you can trust.
Ma.ai does not modify The Crusible's internal reasoning. Four structured stages run inside a sovereign external engine. Ma.ai controls what goes in — not what happens inside.
- Map — entities, relationships, prior decisions
- Eliminate — obvious answers removed first
- Pressure-test — assumptions stress-tested
- Reveal — inferences explicitly labelled
Audit trail & data governance.
Every decision is declared, documented, and defensible. Your data never leaves your control.
You own the record.
Unlike generic AI tools, Ma.ai produces a signed, append-only Correlation Documentation Record for every decision. Exportable to your auditors, your board, and your legal team. The limitation disclosure is hardcoded — never generated by AI.
- Append-only failure library
- Immutable context packages
- Hardcoded limitation disclosure
- Org-scoped RLS on all data
Append-only audit log
Every decision, assumption, and Revelation is immutable. No UPDATE or DELETE on the failure library.
Org-scoped data isolation
Row-level security on every table. Your graph nodes are never accessible by another organisation.
Declared context gate
Human-in-the-loop confirmation before every API call. Context is signed and immutable after confirmation.
Hardcoded limitation disclosure
Every output carries the disclosure as an app-layer constant. Never AI-generated. Never removable.
Board-ready narrative
Operational Narrative generated alongside the Correlation Record. Compliance-ready on day one.
AES-256 encryption
Prompt encryption at rest. Supabase service role key never exposed client-side.
Your organisation's
knowledge, mapped.
Every Crusible run seeds the graph with new entities and relationships. Every failure gets logged. The graph thickens with every decision — making the next Revelation sharper than the last.
- People, companies, contracts, decisions, risks
- Relationships typed and versioned
- Failure library feeds back automatically
- Ontology mode grounds Crusible in your history
Ready to examine
your decisions properly?
Ma.ai is in closed access. Approved organisations get immediate access to Naked mode — direct Crusible runs with full audit trail. Ontology mode unlocks on Phase 2.
- No unexamined vendor approvals
- Every decision defensible to auditors and boards
- Correlation record from day one
Powered by The Crusible · thecrusible.systems