Vérifiez votre IA fournisseur, modèle, conformité, risques, sous-traitants, transparencefournisseurmodèleconformitérisquessous-traitantstransparencefournisseur

Autonoma is an AI-powered supplier discovery and procurement platform. Users describe a sourcing need in plain language (e.g. 'cloud ERP vendors, SOC 2 compliant, in North America') and AI agents search the web to find and enrich verified suppliers with company data, contact emails, relevance scores and trust signals. It then drafts personalized RFI messages, runs outreach campaigns via email or contact forms, and scores/tracks supplier proposals. Intended users are procurement teams in mid-market and growth companies, including Heads of Procurement, Strategic Buyers, and CPOs/VP Operations. Evidenced use cases: supplier find & enrich, RFI drafting, outreach campaign management, and proposal scoring/comparison.

Open medical large language model developed by EPFL and partners (LAION/meditron project) for clinical reasoning. Qantev's official blog (automated medical coding series) documents that its medical-coding methodology leverages foundational LLMs like Meditron, enhanced with prompt engineering (Chain-of-Thought, Few-Shot) and Supervised Fine-Tuning for ICD-10 medical code inference. Note: TechCrunch reports Qantev's production product relies on proprietary small specialized models rather than a single large public model, so this reflects documented methodology/research use.

Open-weight large language model family developed by Meta. Qantev's official blog (automated medical coding series) documents that its medical-coding methodology leverages foundational LLMs like Llama 2, enhanced with prompt engineering (Chain-of-Thought, Few-Shot) and Supervised Fine-Tuning for ICD-10 medical code inference. Note: TechCrunch reports Qantev's production product relies on proprietary small specialized models rather than a single large public model, so this reflects documented methodology/research use.

Fabric is an AI-powered workspace and personal knowledge platform ('your second brain') that unifies files, notes, photos, videos and meeting recordings in one searchable cloud drive with an embedded AI assistant. Fabric mounts as a drive on the user's computer and streams files on demand across web, desktop and mobile, while AI agents (e.g. Sol, Juno, Pip, Mochi, Ada) search, summarize, draft and act across the user's connected apps. Intended users include researchers (including user researchers who keep interviews, transcripts and synthesis notes searchable by meaning), founders, designers, marketers, students and freelancers. It offers deep semantic/visual/color/video/audio search, meeting transcription and auto-summary, business/project planning tools, client publishing portals, and 50+ integrations including Google Drive, Notion, Slack, Figma and GitHub.

Superlegal is an AI-native contract review and contract lifecycle platform, marketed as an 'AI law firm for builders', that combines AI-powered automated contract review, redlining and obligation tracking with expert commercial attorney oversight. It targets high-growth companies and construction/property businesses, providing contract review with a documented 24-hour turnaround, market-standard redlines based on custom playbooks, centralized contract workspace with automatic key-term extraction and obligation/deadline tracking, and legal accountability via attorney review of every contract. Intended users are legal, procurement, and operations teams at construction, technology, marketing, and life sciences/healthcare companies, particularly those without in-house legal. Evidenced use cases include review of prime contracts, subcontractor agreements, purchase orders, MSAs, SLAs, DPAs, NDAs and software licensing agreements, with vendor claims of 70% faster deal velocity and up to 90% lower legal cost.

Trunk Tools' proprietary construction intelligence model family. Trained on 5+ years of real jobsite data (more than 2 million labels from drawings) and 400+ object types. Includes a vision layer that reads drawings (geometry, symbols, spatial relationships, cross-sheet references, revision clouds) plus text/document understanding for specs, submittals, RFIs, schedules, contracts. No specific base model or version publicly documented.