Governed decision infrastructure for enterprises that cannot afford silent failures, invisible risk, or unaccountable AI.
Website | Platform Repository | Investor Thesis | Investor Dashboard
Current status: active prototype / demo platform under development.
No material action is designed to execute without human approval. Demo connectors use mock data unless configured otherwise.
SZL Holdings builds the governed decision infrastructure layer — the platform that connects what is observable to what is executable, under governance, with full attribution.
The accountability gap in enterprise operations is not a data problem. Organizations have dashboards. They have alerts. They lack a system that answers: What do we do next? Who authorized it? What happened as a result? Can we prove it?
We build that system.
A11oy is the flagship governed agentic execution fabric. It senses, structures, correlates, explains, recommends, approves, executes, verifies, and preserves cryptographic proof — in real time, across all SZL verticals.
Every step is instrumented. Every decision is attributed. No material action executes without human confirmation.
Signal → Context → Recommendation → Simulation → Policy → Approval → Execution → Proof → Outcome
A11oy surfaces:
| Surface | What It Provides |
|---|---|
| Now Board | Real-time prioritized signal stream across all verticals |
| Command Surface | Unified operator view — all pending decisions in one place |
| Action Rail | Human-approval queue — no action executes without confirmation |
| Covenant Governance | Policy enforcement — structural, not optional |
| Proof Ledger | Cryptographic audit trail — every decision, immutable |
| Boardroom Mode | Executive decision briefing with full attribution |
| Operator Control Plane | Agent orchestration with observable execution trace |
| MirrorEval | AI reasoning evaluation — closed-loop accuracy tracking |
| Twin Foundry | Digital twin simulation — confidence intervals before action |
| Connector Firewall | Governed integration layer — policy-gated data access |
Live at: /a11oy/
AI governance is structural, not policy-based.
Agents cannot execute consequential actions without human confirmation — enforced at the A11oy governance layer. There is no UI toggle to bypass it.
Evidence-backed decisions.
Every AI recommendation includes source citations, confidence scores, and retrieval provenance. No opaque outputs.
Immutable audit trail.
Every action, approval, and AI recommendation generates an append-only Proof Ledger entry with full decision provenance.
Signal-to-outcome traceability.
From raw signal through routing logic through human approval through executed outcome — every step is logged, linked, and replayable. Zero black-box decisions.
| Product | Domain | Status |
|---|---|---|
| A11oy | Governed agentic execution fabric — signal-to-proof across all verticals | Active — Phase 1 |
| Lyte | Flagship command surface — PRAXIS framework (People, Revenue, Infrastructure, Security, Market) | Active |
| Pulse | AI executive briefing — daily intelligence digest across every domain, source-cited | Active |
| Sentra | Cyber resilience command — exposure mapping, recovery readiness, incident command | Active |
| Vessels | Maritime fleet command — AIS tracking, sanctions screening, voyage economics | Active |
| Terra | Real estate intelligence — distress signals, ownership graph, deal pipeline | Active |
| Counsel | Legal matter command — diligence, evidence, and AI-assisted contract risk | Active |
| Carlota Jo | Premium advisory operations for UHNW clients | Active |
| Aegis | Defense and intelligence command — SOC command, SOAR playbooks, threat intelligence | Active |
| Mobile Command | Unified mobile command — all domain workspaces in one Expo/React Native app | Deferred |
External Signals (integrations, telemetry, intelligence feeds)
|
v
Signal Normalization (A11oy)
|
v
Context Engine (correlation, attribution, scoring)
|
v
Routing (priority, role assignment, domain)
/ \
v v
Auto-Execute Human Review Gate
(policy-approved) |
\ /
v v
Action Execution
|
v
Immutable Proof Ledger (cryptographic, append-only)
|
v
Outcome Graph (closed-loop learning)
+------------------------------------------------------------------------+
| SZL HOLDINGS |
| Governed decision infrastructure layer |
+------------------------------------------------------------------------+
| A11oy — Governed agentic execution fabric (signal-to-proof) |
+------------------------------------------------------------------------+
| Lyte — Command surface (PRAXIS framework) |
| Command — Unified operator surface (cross-domain) |
| Mobile Command — Unified mobile command (iOS + Android) |
+------------------------------------------------------------------------+
| DOMAIN PACKS |
| |
| Sentra Counsel Aegis Vessels Terra |
| Carlota Jo Pulse |
+------------------------------------------------------------------------+
| GOVERNANCE INFRASTRUCTURE |
| |
| Outcome Graph . Proof Chain . Covenant Policy |
| Decision Simulation . Workflow Engine . Event Fabric |
+------------------------------------------------------------------------+
| DATA LAYER |
| |
| PostgreSQL 16 (Drizzle ORM) . External feeds: AIS, STIX/TAXII |
+------------------------------------------------------------------------+
| Concern | Structural Response |
|---|---|
| AI without oversight | Advisory agents require explicit human confirmation — enforced at the A11oy governance layer |
| Opaque AI outputs | All recommendations include source citations, confidence scores, and retrieval provenance |
| Audit accountability | Every action generates an immutable Proof Ledger entry with actor attribution |
| Access control | 11-role RBAC with org-scoped tenant isolation, deny-by-default enforcement |
| Multi-tenancy | All queries scoped by org_id — cross-tenant access architecturally prevented |
| Data security | TLS 1.3, HMAC-signed WebSocket tickets (5-min TTL), encryption at rest |
No SOC 2, HIPAA, FedRAMP, or ISO certifications claimed. Compliance readiness is a planned roadmap item.
Verified, unmodified captures from the active demo platform. No mockups or AI-generated imagery.
A11oy Hero — Seven-Layer Governed Fabric
Captured 2026-04-25. Seven-layer governed agentic fabric with live in-memory seed data.
Boardroom Mode — Executive Governed Decision Briefing
Boardroom Mode — executive decision surface with full attribution and proof chain links.
Command Surface — Unified Operator View
Cross-vertical operator command with human-approval queue. No action executes without confirmation.
SZL Holdings Dashboard — Cross-Portfolio Command
Lyte — PRAXIS Command Surface
Command — Unified Executive View
Sentra — SOC Command
Vessels — Fleet Command
Terra — Deal Pipeline
Carlota Jo — Client Portal
A11oy screenshots captured 2026-04-25 from live server. Platform surface screenshots captured 2026-04-21. Full manifest:
docs/assets/screenshots/current/screenshot-manifest.md(81 A11oy captures across 19 routes and 5 viewports).
→ szl-holdings/szl-holdings-platform
The canonical platform monorepo. TypeScript throughout. A11oy, all domain packs, API server, and documentation live here.
| Item | Status |
|---|---|
| A11oy Phase 1 — Foundation (type system, fabric, demo seed, read API) | ✅ Complete |
| A11oy Phase 2 — Workcell engine with live AI reasoning | 🔜 Planned |
| A11oy Phase 3 — Full proof-carrying execution, live connectors | 🔜 Planned |
| Mobile Command (iOS + Android unified command) | 🔜 Deferred |
| SOC 2 Type 1 audit readiness | 🔜 Roadmap |
| Production customer onboarding | 🔜 Roadmap |
Stephen Lutar — Founder & CEO
Website: szlholdings.com
LinkedIn: linkedin.com/in/stephen-l-279315240
Enterprise & Design Partner: inquiries@szlholdings.com
Investment Conversations: stephen@szlholdings.com
Open to design partner conversations, enterprise evaluation, and investment introductions.