public
Evidence visible on a public Geniuspace URL or downloadable public asset.
Chaque promesse produit est reliée à un niveau de preuve, une limite explicite et une prochaine action de validation. Objectif : rendre Geniuspace lisible par les décideurs, Google et les agents IA sans créer de claim fragile.
Evidence visible on a public Geniuspace URL or downloadable public asset.
Functionality visible as a public prototype or demo; production scope may be narrower.
Internal documentation or filing exists but public proof is incomplete.
External/public registry, benchmark, audit or third-party validation should be attached before using the claim commercially.
Cette page doit être utilisée comme garde-fou éditorial : un claim commercial fort doit pointer vers une preuve, une portée et une limite. Si la preuve externe manque, le claim reste utilisable comme vision produit, pas comme preuve définitive.
| Zone | Claim | Statut | Preuves / sources | Limite explicite | Prochaine action |
|---|---|---|---|---|---|
| Reference Engine | Geniuspace exposes a Reference Engine with hubs, tools, lexicons and localized decision pages. | public |
/fr/reference-engine/ /fr/outils/ /fr/lexique/ /zh/reference-engine/ |
Public pages prove information architecture and UX, not necessarily enterprise adoption. | Add anonymized usage metrics and case-study outcomes when available. |
| Privacy / UX | Several commercial calculators and diagnostic tools can run client-side without collecting personal data by default. | public |
/fr/outils/ /assets/js/reference-engine-tools.js |
A page-level claim does not cover every future form, CRM integration or external analytics tag. | Maintain a tool-by-tool privacy table in the Trust Center. |
| GeniusBrain | GeniusBrain Lite is available as a limited public demo for commercial decision support. | prototype |
/fr/geniusbrain-demo/ |
The Lite demo must not be presented as the complete production model or as a validated autonomous negotiator. | Publish a model card describing scope, inputs, outputs and failure cases. |
| Model architecture | GeniusBrain has an internal documented model-architecture target; public parameter-count claims require a model card and benchmark sheet. | internal |
Internal architecture notes; public summary in llms.txt |
Do not use hard parameter counts as proof until parameter accounting, reproducible build notes, latency, memory and failure cases are published. | Publish model card v1 with parameter accounting, hardware targets, latency and memory benchmarks. |
| Privacy / Security | The product direction is privacy-by-design and favors on-device or local-first processing where possible. | prototype |
/fr/trust-center/ /fr/geniusbrain-demo/ |
On-device claims apply only to components explicitly documented as local; enterprise deployments may add server-side workflows. | Add a data-flow diagram by product module. |
| Intellectual Property | Geniuspace documents an internal portfolio of invention concepts around intercultural behavioral AI, DÉCLIC OS, CCE, ProofCraft and claim governance. | internal |
Internal IP documentation; protected architecture documents |
Do not use patent-pending wording publicly unless application numbers, filing dates and jurisdictions can be disclosed. | Replace broad claims with a public IP table when references can be disclosed. |
| Brand | Geniuspace® is used as the project and brand mark. | external-needed |
Brand usage across public pages and assets |
Trademark registry details should be linked to the relevant official record before claiming registered status in sales material. | Add INPI / EUIPO reference number and scope if public. |
| Product | DÉCLIC OS is presented as a commercial cockpit with scorecards, next-best-actions and report exports. | prototype |
/fr/declic-os/ |
The public page demonstrates product direction; it does not prove multi-tenant SaaS maturity. | Add release stage, screenshots, access model and roadmap. |
| Intercultural B2B | Geniuspace provides frameworks for Japan, China, France and US-oriented B2B negotiation contexts. | public |
/fr/negociation-interculturelle-b2b/ /ja/nemawashi-b2b/ /zh/guanxi-b2b/ |
Frameworks should be localized or reviewed for sector-specific, legal and regional differences before operational use. | Attach expert review notes or partner validation by market. |
| SEO / International | The site exposes French, English, Simplified Chinese and Japanese URL families with canonical and hreflang signals. | public |
/sitemap.xml Page-level canonical and hreflang tags on localized public pages |
Private JSON data is server-side only; crawlers primarily rely on sitemap and page-level hreflang. | Run scheduled Search Console checks after deployment. |
| GeniusBrain | GeniusBrain Local Core v5 demonstrates a local deterministic intercultural decision-support engine with CTE, graph scoring, DÉCLIC calibration, ProofCraft and audit traces. | prototype |
/fr/geniusbrain-demo/ /assets/js/geniusbrain-local-core.js /_data/geniusbrain_protocol_v5.json |
This proves a browser-side prototype and invention architecture, not a complete production model, benchmarked neural network or granted patent. | Attach model card, device benchmarks, test vectors and legal filing references when publishable. |
| DÉCLIC OS / CCE / ProofCraft | DÉCLIC OS v5 exposes a culturally calibrated score engine and ProofCraft before/after proof-normalization report. | prototype |
/fr/declic-os/ /_data/declic_cce_proofcraft_protocol_v5.json /_engine/GeniusBrainEngine.php |
Prototype demonstrates calculation logic; it does not prove SaaS maturity, customer outcome or regulatory certification. | Publish anonymized pilot evidence and before/after ProofIndex case studies. |
| Trust / Claim Governance | EvidenceClaimGate provides a software mechanism for downgrading or sanitizing strong public claims when evidence status is insufficient. | prototype |
/_engine/EvidenceClaimGate.php /_data/evidence_matrix.json |
Server integration must be enforced in all page-generation paths before claiming full automated governance. | Wire claim gating into admin publishing, sitemap/llms generation and CI checks. |
| GeniusBrain | GeniusBrain v6 exposes a local deterministic prototype demonstrating CTE, CGA, Trust Gap Vector, DÉCLIC state machine, ProofCraft, Evidence Ledger and agent-ready output. | prototype |
/fr/geniusbrain-demo/ /assets/js/geniusbrain-local-core.js /_data/geniusbrain_protocol_v6.json |
Prototype demonstration only; not proof of model accuracy, patent filing, production compliance or a released trained model. | Add model card, benchmark report and independent prior-art review. |
| Cultural Graph Attention | A 98-node cultural knowledge graph from the previous project was imported as an internal technical seed for CGA. | internal |
/_data/cultural_knowledge_graph_v6.json /_protected/imported-from-deploy/public/data/CKG_FULL.json |
Internal seed; cultural validity and bias review are required before public performance claims. | Run cultural expert review and benchmark. |
| ProofCraft | ProofCraft v6 defines a hash-linked Evidence Ledger schema for before/after proof and recommendation traceability. | prototype |
/fr/evidence-ledger/ /_data/evidence_ledger_schema_v6.json /_engine/GeniusBrainV6Engine.php |
Browser demo uses lightweight hashing; production should use cryptographic SHA-256 and customer-controlled storage. | Implement IndexedDB/server ledger with crypto-grade hashing and tamper tests. |
| Agent-ready Procurement | GeniusBrain v6 defines an agent-ready procurement packet linking offer, proof, limitations, privacy boundary, cultural calibration and ledger hash. | prototype |
/fr/agent-ready-procurement/ /_data/agent_ready_procurement_schema_v6.json |
Schema and demo output only; no claim of adoption by buyer agents or procurement platforms. | Create JSON-LD examples and validation tests. |
| DÉCLIC OS | DÉCLIC OS v6 is represented as a state machine with observable transition conditions and evidence requirements. | prototype |
/fr/declic-os/ /_data/declic_state_machine_v6.json /_engine/GeniusBrainV6Engine.php |
State machine is deterministic and needs field validation before effectiveness claims. | Benchmark against annotated commercial opportunity histories. |
| ProofCraft | ProofCraft exposes a public Score Card v2 prototype with Proof Index and five KPIs: L3R, NSCR, DMC, ED and CAR. | prototype |
/fr/proofcraft/ /_data/proofcraft_protocol_v2_2026_2027.json /assets/js/reference-engine-tools.js |
Indicative scoring prototype; field effectiveness requires customer outcome benchmarks and independent validation. | Attach anonymized 14-day sprint outcomes and publish a model/metric card. |
| GeniusEngine v7 | GeniusEngine exposes six public prototype modules: GeniusTwin, Trust Gap Vector, Proof Compiler, Behavioral OS Layer, Intercultural Alpha Score and Agentic Negotiation Arena. | prototype |
/fr/cultural-digital-twin/ /fr/trust-gap-vector/ /fr/proof-compiler/ /fr/behavioral-os-layer/ /fr/intercultural-alpha-score/ /fr/agentic-negotiation-arena/ /assets/js/reference-engine-tools.js |
The modules are deterministic local prototypes, not validated autonomous negotiators or certified predictors. | Publish module cards, failure cases and anonymized outcome benchmarks. |
| ProofCraft / Evidence Ledger | Proof Compiler can generate a local Deal Passport prototype with KPI scorecard, cultural signature, event list and hash. | prototype |
/fr/proof-compiler/ /fr/evidence-ledger/ /fr/proofcraft/ |
Production deployment should use cryptographic SHA-256, customer-controlled storage, sourceRefs and human validation. | Add an exportable JSON schema and customer-controlled ledger sample. |
| Commercial metrics | Intercultural Alpha Score provides a prototype before/after metric for cultural calibration impact. | prototype |
/fr/intercultural-alpha-score/ |
Marketing averages must not be stated until backed by real client outcome data. | Collect anonymized 14-day Proof Sprint baselines and outcomes. |
| GeniusEngine v9 uses a public CTE/CGA-inspired runtime and Cultural Knowledge Graph for deterministic prototype scoring. | prototype |
/fr/cultural-knowledge-graph/ assets/data/CKG_FULL.json _data/geniusengine_v9_cga_manifest.json |
|||
| Reference Engine v10 relie claims, preuves, sources, versions et modules dans un système agent-readable. | implemented_static_prototype |
/fr/reference-engine-intelligence/ /assets/data/reference_graph_v10.json /reference.json /.well-known/geniuspace-reference.json |
|||
| Reference Engine / Tools | Geniuspace exposes a local-first Tools Intelligence Layer with commercial calculators, simulators and audits for retail, IA sales, negotiation, China GTM and AI SEO/GEO. | public |
/fr/outils-commerciaux/ /reference/tools.json /assets/data/tools_intelligence_v11.json |
Scores are deterministic heuristics and must not be presented as audited results or guaranteed business outcomes. | |
| GeniusBrain local AI | GeniusBrain exposes a browser-native WASM/WebGPU prototype for local intercultural scoring. | prototype |
/fr/geniusbrain-wasm-webgpu/ /assets/wasm/geniusbrain_cte_cga_v7.wasm /assets/js/geniusbrain-wasm-webgpu-v7.js |
The WASM kernel is deterministic and small; ONNX Runtime WebGPU is an optional adapter, not a bundled production ML model. | Bundle validated ONNX models and publish reproducible benchmark results. |
| Agent-readable procurement | Geniuspace can generate JSON-LD procurement packets containing Offer, TrustGapVector, CulturalFitScore and ProofCompiler output. | public |
/fr/agent-readable-offer-engine/ /reference/offers.json /.well-known/agent-readable-offer-engine.json |
Agent ingestion by third-party procurement bots is not guaranteed. | Test with enterprise procurement workflows and publish parser compatibility. |
| Evidence Ledger | Evidence Ledger can emit W3C-style Proof Sprint Verifiable Credential JSON for verification workflows. | prototype |
/fr/evidence-ledger-verifiable-credentials/ /reference/evidence-ledger-vc.json |
Demo credentials require production key management, revocation registry and verifier implementation. | Implement real signing service and verification endpoint. |
| Benchmarks | AKST provides a public seed benchmark for adversarial cultural stress tests in sales AI. | public |
/fr/adversarial-cultural-stress-test/ /benchmarks/akst_v12.json |
Initial scenarios are seed tests; cultural validity requires expert review and field evidence. | Add peer-reviewed cases and scoring reliability metrics. |
Peuvent être utilisés dans les titres, meta descriptions et pages d’offre si la preuve est accessible depuis une URL publique.
À formuler comme démonstration, aperçu ou version Lite. Ne pas promettre un périmètre enterprise tant qu’il n’est pas documenté.
À garder cadrés : préciser “en cours”, “documenté en interne” ou “référence publique à venir” tant que les numéros et audits ne sont pas publiés.