CogniAI
Team consisting of a Peregrine.ai AI engineer (MSc Data Science) skilled in PyTorch, Rust/Substrate, Solidity; an ESCP/Mines IoT designer; and a 42 Paris low‑code builder.
Project Description
Learning today has paradoxically become difficult. Traditional educational systems are rigid and do not adapt to individual learning styles, levels, or pacing. Content is abundant but unstructured, and current platforms—YouTube, Udemy, MOOCs—do not provide truly personalized guidance. Learners often lose motivation due to lack of self-assessment tools, with 70% abandoning a MOOC before completion. In professional settings, companies lack time and resources to build tailored onboarding or training programs, with 50% reporting that internal trainings are ineffective. This gap in adaptation, structure, and guidance represents a major barrier to effective learning.
CognIA addresses this problem with a fully autonomous AI learning companion for students, professionals, and self-learners. Using a fleet of specialized AI agents, CognIA assesses each learner’s level via quizzes or questionnaires and creates a custom curriculum by combining resources from the web, internal documents, books, and videos. The learning path adapts in real time to the user’s pace, style, and preferences. CognIA automatically generates lessons, flashcards, exercises, projects, summaries, and mock exams while providing motivation, guidance, and day-to-day coaching. A cognitive twin continuously learns the user’s way of learning, tailoring the curriculum like a personal tutor who knows them perfectly.
CognIA is powered by five autonomous agents for the :
Explorer agent (content miner) identifies resources, Archeologist (content analyser) analyzes and structures them, Librarian (database manager) organizes the library and generates learning blocks, Composer (curriculum architect) **builds and adjusts personalized curricula, **Guide (Coach mentor) interacts with and motivates learners.
Other agents drive autonomously company operations : Growth Captain (marketing agent) manages marketing and user acquisition, Caretaker (support and success) handles 100% AI customer support, and Strategist (product analyst) continuously optimizes the product, curricula, and engagement. This architecture allows CognIA to operate almost entirely independently, meeting the hackathon goal of a “One-Person Unicorn” where agents handle research, operations, growth, and support.
Value proposition: advanced personalization, active engagement through note-taking and interactive feedback, and full autonomy for learners to create and enhance their own path.
Key features: intelligent diagnostics, dynamic curriculum, automated content generation, import of PDFs/videos/documents, personalized mock exams, smart note-taking, and active learning.
Tools used from partners:
- Google Cloud, Mistral (LLM), Figma make, lovable, n8n
other tools : Cursor, CrewAI, Northflank, Supabase, Qonto, Stripe, Notion
Business Model
-
Freemium:
- Free version with core features (curriculum creation, simple exercises).
- Premium version with access to exclusive databases, advanced AI mentoring, and detailed analytics.
-
Partnerships:
- Collaboration with schools, universities, or companies to deliver customized versions.
-
Subscription:
- Monthly or yearly plans for individuals and professionals.
3 Market Segments
1) Education (large-scale)
- primary to high school students
- university students
- private schools
- tutoring and academic support
2) Enterprise : automated onboarding
- internal training
- rapid upskilling
- compliance and regulatory training
3) General Public
- languages
- computer science
- creative skills
- personal development