LMS (Learning Management System)
Platform focused on training administration: enrollment, tracking, reporting. Historically built for HR teams more than learners. Often perceived as rigid and austere.
Glossary
LXP, LMS, RAG, embeddings, AI agents, vectorization… terms that come up constantly. Here are our definitions, written for L&D and People teams who want to understand without jargon, with concrete examples anchored in corporate learning.
Platform focused on training administration: enrollment, tracking, reporting. Historically built for HR teams more than learners. Often perceived as rigid and austere.
New generation of platforms focused on learner experience: recommendations, adaptive paths, gamification, short content. Hopra goes further with native AI integrated at every layer.
Client instance of the platform. In a multi-tenant architecture like Hopra, each tenant has its own branding, content, users, with strict data isolation. A single organization can run multiple tenants (by business unit, brand, geography).
Architecture where a single codebase serves N simultaneous clients, each isolated from the others. Enables fast deployments (48h at Hopra) and upgrades without migration, with massive infrastructure economies of scale.
Commercial model where a product is resold under the client's brand. Hopra offers full white-label: web branding, mobile app, communications, with no editor mentions.
Training of end users of a product or service. Becomes a strategic lever for SaaS editors: reduces churn, accelerates adoption, creates product ambassadors.
Platform where AI is the system's architecture, not an added feature. Every interaction flows through the AI engine. As opposed to AI "glued onto" a legacy tool.
Autonomous program powered by an LLM, specialized for a precise task. An enterprise AI agent can read documents, synthesize, create, coach. Hopra ships 4 specialized agents: Creation, Coach, Knowledge, Analytics.
Technique that grounds AI answers in a trusted document base. The AI retrieves relevant passages (retrieval) before generating an answer (generation) with source citations. Zero invention, everything is traceable.
Numerical representation of a text's meaning, as a vector. Two passages about the same idea will have close embeddings, even using different words. This is what enables semantic search.
Process of turning content (text, image, audio) into an embedding. At Hopra, your knowledge base is vectorized continuously to power semantic search and RAG.
Search that understands the meaning of a query, not just keywords. "How to handle burnout" will find a document on professional exhaustion, even without the word "burnout".
Massive language model (GPT, Claude, etc.). LLMs alone hallucinate easily. At Hopra, they're systematically framed by RAG on your base to guarantee traceable answers.
When a language model invents an answer that sounds right but is false or made-up. RAG and grounding in your sources drastically reduce this risk.
AI assistant embedded in an application to continuously help the user. Hopra offers a learner copilot (conversational coach) and a trainer copilot (pedagogical creation assistant).
Refining a language model on specific data to specialize it. Expensive and rarely needed when RAG is enough — Hopra prefers RAG to stay agile and protect your data.
Mandatory French certification for training organizations wanting access to public funds (CPF, OPCO). Requires strict traceability of every learning path. Hopra industrializes this traceability without extra friction.
European regulation on personal data protection. Mandates EU hosting, right to erasure, portability, explicit consent. At Hopra, compliance by design — European hosting, zero third-party sharing.
Enterprise single sign-on protocols. SAML 2.0 for legacy Microsoft ecosystems, OIDC for modern systems. Both supported by Hopra with SCIM provisioning.
Pedagogical content packaging standard (1.2, 2004). Still omnipresent in corporate catalogs. Hopra reads and writes SCORM to interoperate with your existing content.
Modern pedagogical tracking standard, more granular than SCORM. Tracks learning activities beyond a single course (mobile, experiential, social learning). Hopra is xAPI-native.
Integration standard between an LMS/LXP and third-party pedagogical tools. Lets you embed an external tool in a path without auth break. Supported by Hopra.
Short pedagogical format (3-8 minutes) designed for mobile consumption and in-between moments. At Hopra, micro-learning capsules are AI-generated from your source documents.
Approach where the path adjusts in real time to each learner's level, pace and preferences. Hopra's native AI makes adaptive learning accessible without complex setup.
Percentage of learners who complete a path. Industry median: 15-30%. With Hopra's premium experience + AI, our pilots see × 2.4 vs legacy LMS.
Group of learners following a path together, with collective dynamics (leaderboards, challenges, rituals). Hopra's manager layer lets you run multiple cohorts in parallel.
Intensive path over a short period (1 to 4 weeks) with sustained pace and precise objectives. Hopra lets managers launch bootcamps for their team in a few clicks.
Integration of game mechanics (XP, streaks, badges, tournaments) in learning. At Hopra, gamification is architectural: adaptive progression map, seasons, 110+ badges.
Subtle indicator of a risk or opportunity detected before the trend becomes obvious. Hopra's AI agents detect drop-off weak signals to allow early intervention.
Let's talk
15 minutes, a concrete case, sharp answers. No pitch, no sales rep.
We reply within 24 business hours. Your data stays in Europe.