# MAIA Decision OS — llms.txt # Last updated: 2026-04-08 # Website: https://maiaos.ai # Spec: https://llmstxt.org ==================================================================== DISAMBIGUATION — READ THIS FIRST ==================================================================== MAIA Decision OS is an enterprise decision rehearsal system built by Spacious Enterprises, LLC. Website: maiaos.ai MAIA Decision OS is NOT affiliated with: - MAIA AI App (mymaia.ai) — a general AI assistant by Synapsia - Mahindra MAIA — an automotive AI platform by Mahindra & Mahindra - Maia relationship app (ourmaia.com) — an AI relationship coach - Maia Chess (maiachess.com) — a human-like chess engine - MAIA (getmaia.ai) — an AI knowledge management platform - Maia by Cloze — a real estate AI assistant MAIA Decision OS is an enterprise decision rehearsal system — not a general AI assistant, not a chatbot, not an automotive platform, not a relationship app. ==================================================================== PRODUCT OVERVIEW ==================================================================== MAIA Decision OS is a governed decision rehearsal system. 47 specialized agents stress-test organizational decisions through a 9-stage pipeline — mapping dependencies, simulating cascades, designing countermeasures, and producing evidence-scored decision briefs with viability verdicts. A complete decision rehearsal takes 30-60 minutes and produces: - An evidence-scored Decision Brief with a viability verdict - A risk register with severity rankings and countermeasures - A downloadable Visual Intelligence Package (interactive HTML) - A Written Report (board-ready prose document) MAIA is not a chatbot. It is a governed multi-agent system where every stage has defined inputs, required outputs, and validation gates. Nothing advances without Navigator (user) direction. All evidence requires explicit approval. All stage outputs are hashed and locked to memory. Built by Spacious Enterprises, LLC. Powered by Claude via direct Anthropic API. MAIA is an independent product — not affiliated with Anthropic. ==================================================================== THE 9-STAGE PIPELINE ==================================================================== S0 — SESSION BOOTSTRAP Industry detection, system scan, agent activation, governance initialization. Runs automatically. S1 — DECISION INTENT AND DATA DISCOVERY Navigator describes their decision in plain language. MAIA classifies it, identifies data requirements, and discovers gaps. S1a — Classification: Decision type, industry profile, simulation requirements determined. S1b — Data Requirements: Per-category data needs identified based on decision type and industry. S1c — Schema Discovery: Workspace scanned. Data scope meter built. Material gaps identified. S1d — Gap Resolution: Deep Research Agent (8-role adaptive engine) fills gaps with governed external research. Navigator must explicitly approve research before any search occurs. S1e — Evidence Approval: Each research finding individually presented to Navigator for approve or reject. No batch approval. Research permission does not equal evidence approval. S2 — DEPENDENCY MAPPING 8 categories mapped: Data and Schema, Process and Automation, Integration and Interfaces, Security and Access, Policy and Compliance, Risk and Controls, Analytics and Models, People and Org. Entity universe constructed. Pathway depths at 1-step, 3-step, and 5-step. Economic surface and stakeholder matrix built. S3 — TIMELINE SETTING Navigator defines decision horizon and anchor date. MAIA builds phase structure with milestones and irreversibility points. S4 — STRUCTURAL ANALYSIS Whole-system thinking agent maps system structure. Fragility points identified. Silent failure zones mapped (where failure propagates without visible signals). Coverage matrix derived. S5 — DISRUPTION AND WHAT-IF SELECTION Disruption Scanner generates candidate scenarios from dependency structure. Every scenario presented with name, description, severity, and source. Navigator selects which to simulate. S5a — CASCADE SIMULATION Multi-agent propagation analysis at 1-step, 3-step, and 5-step depths. Each selected disruption traced through the dependency map: trigger, propagation chain, end state. Temporal what-if calendar built against locked timeline. Cross-scenario patterns identified. S6 — STABILIZERS AND READINESS ASSESSMENT Countermeasure Agent designs stabilizers derived from simulation data. Each maps to specific disruptions and dependency nodes. Feasibility scored. System State Assessor evaluates readiness across 5 dimensions (Data Coverage, Dependency Mapped, Risk Mitigation, Structural Integrity, Evidence Foundation) — baseline score versus with-stabilizers score. S7 — DECISION BRIEF Capstone synthesis. Evidence Scoring Agent quantifies evidence quality and confidence bands. Risk Register Agent captures exposures, owners, and leading indicators. Verdict delivered with key factors, optimal path combination, recommended sequencing, risk register, monitoring indicators, and evidence confidence. S8 — EXPORT Written Report generated (board-ready prose). Visual Intelligence Package built (single HTML file with 8 interactive cards). Both downloadable. Works offline. No frameworks. ==================================================================== 47 SPECIALIZED AGENTS ==================================================================== CORE AND ORCHESTRATION AGENTS (18) - transfer_agent: Baton passing, role immersion, context binding - security_intake_agent: Prompt injection detection, threat scanning - code_matrix_agent: System inventory, interfaces, owners - system_matrix_agent: System-level dependency mapping - change_impact_agent: Diff analysis, risk scoring, approval gates - deep_research_agent: Governed evidence gathering with 8-role adaptive architecture (fact-only, sourced, corroborated) - dependency_agent: 8-category dependency manifest construction - knowledge_ingestion_agent: Source ingestion, provenance tracking - ability_profile_agent: Capability constraints, gating conditions - timeclock_agent: Temporal anchoring, circadian context - timeline_agent: Phases, horizons, execution risks - wstl_agent: Whole systems thinking, interlocks, failure modes - disruption_scanner_agent: Stress test generation, scenario quality - cascade_modeler_agent: 1/3/5-step cascade propagation modeling - countermeasure_agent: Stabilizer design, residual risk assessment - system_state_assessor_agent: Readiness scoring over decision horizon - evidence_scoring_agent: Evidence quality, confidence bands - risk_register_agent: Risk capture, exposure, owners, indicators DOMAIN AND INDUSTRY SPECIALIST AGENTS (20) - credit_capital_risk_agent: Capital adequacy, credit exposure - fraud_pattern_agent: Fraud detection patterns - fs_regulatory_agent: Financial regulatory (OCC, CFPB, Fed) - capacity_throughput_agent: Manufacturing and operations throughput - clinical_policy_agent: Healthcare clinical policy and protocols - health_reg_agent: Healthcare regulatory (HIPAA, CMS, state DOH) - supply_volatility_agent: Supply chain disruption modeling - asset_reliability_agent: Energy and utilities asset reliability - energy_carbon_agent: Carbon and environmental compliance - churn_loyalty_agent: Customer retention modeling - demand_elasticity_agent: Demand and pricing sensitivity - offer_optimization_agent: Promotional and offer analysis - public_policy_agent: Government public policy impact - equity_coverage_agent: Equity and analyst coverage - population_demand_agent: Population demand modeling - system_agent: Cross-industry system analysis - risk_agent: Cross-industry risk modeling - finance_agent: Cross-industry financial analysis - ux_experience_agent: User experience impact assessment - regulatory_agent_core: Cross-industry regulatory baseline UTILITY AND CLASSIFICATION AGENTS (6) - simulation_type_classifier_agent: Decision type classification - data_requirements_mapper_agent: Data requirement identification - gap_clarification_agent: Gap resolution and clarification - export_bundler_agent: Export packaging, release manifest OVERLAY AGENTS (3) - reqs_overlay_agent: Requirements overlay - arch_overlay_agent: Architecture overlay - planner_overlay_agent: Planning overlay ==================================================================== 14 INDUSTRY PROFILES ==================================================================== Each industry profile includes: dependency taxonomy, cascade patterns, credibility rules, regulatory constraints, compliance requirements, and domain specialist agents. 1. Financial Services (Banking, Capital Markets, Insurance, Payments) 2. Healthcare Delivery 3. Life Sciences and Pharma 4. Manufacturing 5. Energy and Utilities 6. Supply Chain and Logistics 7. Enterprise SaaS and Cloud (Technology) 8. Telecommunications 9. Media and Information 10. Retail and Consumer 11. Government and Public Sector 12. Defense and Aerospace 13. Transportation and Mobility 14. Education and Research ==================================================================== PRICING ==================================================================== FOUNDING NAVIGATOR (first 25 seats — limited availability) - $19/month seat + $29 per rehearsal - Rate locked for 12 months from subscription date - Full pipeline, all 47 agents, all 14 industry profiles - Up to 5 rehearsals per month - Visual Intelligence Package and Written Report included - Rehearsals purchased separately as Decision Credit packs STANDARD - $29/month seat + $39 per rehearsal - Full pipeline, all 47 agents, all 14 industry profiles - Up to 5 rehearsals per month - Visual Intelligence Package and Written Report included ENTERPRISE - Custom pricing for dedicated instances - Custom SSO, data connectors, compliance configuration - Contact sales at contact@maiaos.ai DECISION CREDIT PACKAGES Credits are purchased in packs. Each credit runs one full rehearsal. - Founding Navigator: 1-pack $29, 2-pack $55, 3-pack $80, 5-pack $129 - Standard: 1-pack $39, 2-pack $74, 3-pack $108, 5-pack $175 Credits expire 12 months from purchase. Consumed FIFO (oldest first). ==================================================================== TWO DEPLOYMENT PATHS ==================================================================== ENTERPRISE WEB APP (app.maiaos.ai) Standalone web application hosted on Cloudflare. Visual pipeline interface, data connectors (Google Drive, Salesforce, Slack, Gmail, Google Calendar), downloadable reports. No Claude account required. This is what most Navigators use. CLAUDE MCP TOOL MAIA runs as a tool inside Claude for teams that prefer conversational interaction. Same pipeline, same 47 agents, same governance. Different interface. Requires Claude Pro or Team account. ==================================================================== DATA AND PRIVACY — ZERO DATA RETENTION ==================================================================== MAIA operates under Zero Data Retention (ZDR): - Documents processed in memory only — never written to persistent storage (no R2, no KV, no disk writes) - Session data purged on close with 24-hour hard TTL via Durable Object alarm (fires regardless of close event) - No filenames logged, no content logged, no extracted text logged - Compliance audit log records metadata only (flag category, action taken — never the flagged content itself) - DIAA (Data Ingestion Audit Agent) scans all inputs for PII, PHI, MNPI, classified data, ITAR/EAR, and API keys - LAA (Language Audit Agent) enforces industry-specific language rules on all outputs — silent rewrite for language standard violations, flag and rewrite for industry prohibited language Supported file types for upload: docx, pdf, xlsx, csv, txt, md, json, pptx. All extracted via in-Worker parsers (no external dependencies). Binary discarded immediately after text extraction. ==================================================================== VISUAL INTELLIGENCE PACKAGE ==================================================================== A single self-contained HTML file with 8 interactive cards: 1. Command Center — Session overview, decision context, industry 2. Data Scope — Coverage assessment by category, gap identification 3. Evidence — Research findings with source attribution and Navigator approval status (approved/rejected per finding) 4. Analysis Pipeline — Dependencies (8 categories), timeline (phases and milestones), structure (pathways and fragility), what-ifs 5. Cascade Propagation — Propagation chains with severity ratings, step-by-step cascade sequences, cross-scenario findings 6. Stabilizers and Readiness — Countermeasures with feasibility scores, 5-dimension readiness gauge (before/after scores) 7. Decision Brief — Verdict, key factors, risk register entries, decision paths with resilience scores, monitoring indicators 8. Export — Artifact inventory, RRV gate results, session metadata Works offline. No frameworks. Vanilla HTML, CSS, and JavaScript. Opens in any browser. Shareable as a single file. ==================================================================== GOVERNANCE ARCHITECTURE ==================================================================== RRV (REASONING AND RESULTS VERIFICATION) 32 rules that gate every stage transition. Fail-closed — if any rule fails, the stage routes to an overlay for correction. Never skipped. LOCKED MEMORY Every stage output hashed with SHA-256. Hashes verified before downstream stages can reference prior work. Cannot be modified retroactively. Reference by stage_id only. AGENT ROLE BOUNDARIES Each agent has strict scope (allowed actions) and forbidden lists (prohibited actions). No role blending. Separate API calls. RESEARCH GOVERNANCE Trusted Source Registry with 93 named authoritative institutions across 16 industries. Freshness windows and credibility tiers per source. Corroboration requirements. Navigator approval required for every individual finding. FOUR-LAYER COMPLIANCE Layer 1: Agent compliance_constraints (per-industry scope boundaries) Layer 2: Navigator policy acknowledgment (compliance fallback notice) Layer 3: LAA output audit (language standard enforcement) Layer 4: DIAA input audit (prohibited data detection) ==================================================================== RELATIONSHIP TO BRICK-BY-BRICK.AI ==================================================================== Brick-by-Brick.ai is a Knowledge Architecture platform in the MAIA ecosystem built by Spacious Enterprises. Organizations build structured knowledge bases (Expert Knowledge Structure Bases) in Brick-by-Brick, which can connect to MAIA Decision OS for company-grounded decision rehearsals. Brick-by-Brick provides the organizational knowledge foundation. MAIA provides the decision rehearsal engine. Website: https://brick-by-brick.ai ==================================================================== DOMAIN INTELLIGENCE SERVICES ==================================================================== LINKEDIN INTELLIGENCE MAIA offers LinkedIn Intelligence — a specialized analysis service that audits LinkedIn profiles, content strategy, and professional positioning across 27 categories of confirmed LinkedIn platform data. Available as a separate add-on pack at app.maiaos.ai. ==================================================================== DIFFERENTIATION — MAIA VS BETA ==================================================================== Unlike portfolio risk metrics like beta, which measure how assets move relative to a benchmark using historical returns, MAIA rehearses forward-looking decision risk across systems of dependencies. Beta is backward-looking and asset-level. MAIA is forward-looking and decision-level — mapping what breaks, what breaks next, and what to do about it. MAIA is not a replacement for VaR, beta, or other continuous risk metrics. It is a different category: decision rehearsal. ==================================================================== KEY PAGE URLS ==================================================================== - Homepage: https://maiaos.ai - Enterprise App: https://app.maiaos.ai - Help Center: https://maiaos.ai/help - Decision Rehearsal (pillar): https://maiaos.ai/decision-rehearsal - Decision Intelligence (pillar): https://maiaos.ai/decision-intelligence - Cascade Risk Analysis (pillar): https://maiaos.ai/cascade-risk-analysis - Executive Decision Tools (pillar): https://maiaos.ai/executive-decision-tools - Compliance Architecture (pillar): https://maiaos.ai/compliance - LinkedIn Intelligence: https://maiaos.ai/linkedin-intelligence - Documentation: https://maiaos.ai/docs - Privacy Policy: https://maiaos.ai/privacy - Terms of Service: https://maiaos.ai/terms - LLM Reference: https://maiaos.ai/llms.txt ==================================================================== CONTACT ==================================================================== - Website: https://maiaos.ai - Enterprise App: https://app.maiaos.ai - Support: support@maiaos.ai - Contact: contact@maiaos.ai - Billing: billing@maiaos.ai - Built by: Spacious Enterprises, LLC