Summary - Product: Generative Semantic Fabric (GSF) platform that automatically discovers, labels, and maps metadata from on-prem and cloud systems into an organizational ontology/virtual knowledge graph to make structured data AI-ready and governed. The description below reflects only published/site claims and documentation reviewed. What illumex claims to do (high-level) - Make structured enterprise data "AI-ready" by operating on metadata: automated discovery, labeling, mapping into a semantic/virtual knowledge graph and auto-generated data dictionary. - Enable deterministic, explainable, governance-first GenAI/agent analytics without moving raw data values; the site repeatedly states it uses metadata only. - Target customers: enterprises with regulated data and complex, heterogeneous data estates (examples cited: on-prem ERP, SAP, POS, contact centers, cloud warehouses, BI tools). Core capabilities (as presented on site/docs) - AI-Ready Data: automated discovery, labeling, mapping of metadata; auto-generated data dictionary; tagging including PII/sensitive-data detection, alerts and recommendations. - Augmented AI Governance: business glossary, definitions, metric standardization, duplication/PII flagging, lineage and governance workflows, and auditability features. - Column-level Data Lineage & Impact Analysis: tracing of origins and transformations, root-cause and impact analysis, graph visualizations. - Generative Semantic Fabric: virtual knowledge graph and semantic embeddings encoding data semantics/relationships to support deterministic, non-hallucinating responses for LLM agents (site framing is that the graph/embeddings remove ad-hoc retrieval risks). - GenAI Agent Deployment: SDK and agent orchestration, integrations (mentions Slack, MS Teams via Omni), deterministic data retrieval and an explanation layer that translates prompts to SQL/LLM runtimes (described in docs/marketing). - Productivity & onboarding claims: automated mapping at scale (examples cited: 100k+ assets auto-labeled; millions of metadata points processed in hours) and time-to-value claims such as "start in ~one week" for initial deployment/metadata onboarding. Security, privacy & compliance (published claims) - The site references SOC 2, GDPR, and ISO 27001 as part of enterprise security posture. The company repeatedly states it works with metadata only and does not move raw data values (privacy/security stance as presented). - There is a Privacy & Security section and a Knowledge Base category for privacy_and_security on docs.illumex.ai. Integrations, partners, customers (public mentions) - Partners / platform mentions: Microsoft, Google Cloud, AWS, NVIDIA (appears in press/news and partner mentions). - Customers referenced in press/case anecdotes: Teva, Carson (as public press/case mentions). - Integrations: references to connectors for enterprise sources (SAP, on-prem systems), BI tools like Tableau, and web extensions; no exhaustive public connector list found in the reviewed pages. Public documentation and pages reviewed - illumex.ai/ (home/product overview) - illumex.ai/semantic-layer-platform/ (semantic layer / platform details) - illumex.ai/documentation/ (documentation index) - docs.illumex.ai/ (docs homepage) - docs.illumex.ai/category/privacy_and_security/ (privacy & security KB category) - illumex.ai/request-demo/ (request demo / contact) - illumex.ai/company/ (company / about) - illumex.ai/press/illumex-raises-13m-to-enable-trustworthy-governed-enterprise-genai-with-structured-data/ (press / funding) - illumex.ai/privacy-and-security/ (privacy & security archive article) - illumex.ai/terms-and-conditions/terms-and-conditions/ (terms) Pricing & trials (public findings) - No public pricing pages or published price plans found on the site or docs. - No free trial or self-serve pricing options discovered; the published approach is enterprise/request-demo/contact-sales for pricing and pilot/POC details. Differentiators vs RAG (as presented) - Claims of lower maintenance via automated semantic upkeep vs manual RAG document/corpus maintenance. - Deterministic, explainable outputs through a governed semantic knowledge graph rather than ad-hoc document retrieval. - Security/privacy benefit emphasized by operating on metadata only (as presented on site). Notable published claims / example outcomes (marketing/explanatory content) - Promotional/example outcomes cited on site: "5x AI productivity in first year", automatic mapping of 100k+ assets, millions of metadata points processed in ~11 hours. - Time-to-value statements such as "start in about one week" for initial deployment/metadata onboarding. Gaps / items not publicly available (observed) - No published pricing tiers, license models, or cost examples on public pages. - No publicly downloadable SOC 2 or ISO 27001 audit reports; no full Data Processing Agreement text found on publicly accessible pages (site references a Data Processing Addendum but the specific documents were not publicly available during this review). - No public trial sign-up or self-serve sandbox documented on the reviewed pages. Recommended next steps (if engaging) 1) Request a tailored demo via the request-demo page and share your goals. 2) Ask sales for: pricing models/tiers and TCO examples; security/compliance packages and any audit reports/DPA/data flow diagrams proving metadata-only handling; a complete integration/connector list and any prebuilt adapters for your stack (SAP, Snowflake, Tableau, etc.); references/case studies for similar customers and measured outcomes. 3) Request a short POC/pilot scope (sample metadata scan + sample mapping), expected onboarding timeline and measurable success metrics. 4) Post-demo, review the docs site for technical details on connectors, APIs, plugins, and admin/setup processes. Notes on source fidelity - This summary reproduces claims and content that are published on illumex.ai and docs.illumex.ai pages reviewed. I did not fabricate claims; where the site uses marketing language it is presented as a site claim (e.g., "start in ~one week", "100k+ assets"), and gaps above note missing public artifacts. If you want next actions I can: draft a demo request email, attempt to locate any additional public technical docs or security attestations, or prepare a checklist of demo questions (integration, data residency, audit evidence, pilot scope, SLA, pricing, exit/rollback).