CHATBOOK TUTOR by Chatbook Hikma AI Lab
CHATBOOK TUTOR by Chatbook Hikma AI Lab
From courses
to personalized
educational
tracks
The advent of artificial intelligence technologies enables a fundamental rethinking of the learning process. Standardised courses and static instructional materials are no longer the only instruments at our disposal. An approach grounded in Individualised Educational Trajectories (IET) makes it possible to build full-fledged corporate universities for companies—where learning serves as a risk-management tool, not merely a checkbox to tick off in an LMS.
If a building becomes architecture, then it is art
Corporate learning evolution
In large holding companies, learning often follows the same familiar pattern: regulations buried in PDFs, courses locked inside an LMS, in-person safety briefings, and on-site mentors. HR tracks completion rates. Department heads, however, rarely see whether an employee actually knows how to act in a non-standard situation. Knowledge is scattered across hundreds of documents; the moment a regulation is updated, the entire cycle has to start over.
The rise of artificial intelligence has given birth to a new class of learning solutions: on-premise AI tutors. Deployed on internal servers, they work with your existing PDFs and DOCX files, retain session context, and build personalised learning paths and assessments on the fly. This shifts training from an annual "retake-the-test" formality to a continuous companion—one that fuels career growth and drives margin expansion alike.
The rise of artificial intelligence has given birth to a new class of learning solutions: on-premise AI tutors. Deployed on internal servers, they work with your existing PDFs and DOCX files, retain session context, and build personalised learning paths and assessments on the fly. This shifts training from an annual "retake-the-test" formality to a continuous companion—one that fuels career growth and drives margin expansion alike.
Business Scaling
Onboarding
Assessment
Downtime reduction
Synchronization of branches
Digital transformation
Accident rate reduction
Learning leaders and HR departments tackle a wide range of challenges through training, yet their ultimate objective remains singular: to mitigate managerial risks across every conceivable scenario.
Training beats hiring on cost
Finding a specialist with industry experience and familiarity with your internal processes takes months. Onboarding them—another quarter. Training an existing employee on your updated procedures using an AI tutor? A matter of weeks.
A bad hire for a mid‑ or senior‑level position costs several times more than the entire annual L&D budget. Investing in competencies, by contrast, preserves accumulated expertise within your organisational perimeter—and keeps it there.
24/7
+XXX
Training serves as a tangible signal of the company's commitment to its people. In a climate of acute talent shortages across both industrial sectors and IT, this perception directly translates into lower turnover—and higher retention.
An external specialist departs—and their intimate grasp of your internal workflows departs. A trained employee, by contrast, cements institutional know‑how within the organisation and channels it forward along those very same digital pathways, steadily compounding your intangible assets.pathways, steadily compounding your intangible assets.
+LTV
+ НМА
Chatbook Tutor —
an AI‑powered product line designed to build corporate universities and enterprise‑grade knowledge systems
an AI‑powered product line designed to build corporate universities and enterprise‑grade knowledge systems
Version 1.0 — a mature platform ready for pilot deployment.
Conversational Tutoring Grounded in Your Documents
A unified agentic loop powered by RAG over your enterprise knowledge base: responses drawn exclusively from your internal regulations — never from the "Internet's memory." Supports PDF, DOCX, XLSX, PPTX; cross‑sectional relationships are mapped via GraphRAG / LightRAG.
Mastery Path Learning Trajectories
Step‑by‑step progression through topics with mastery gates: you move forward only after demonstrated understanding. This is not a linear "click‑through" course, but a governed, adaptive route.
Assessments & Question Bank
Automated generation and administration of questions with detailed explanations — designed for reinforcement and self‑checking.
Library of AI Mentors
Persistent mentors with predefined personas and memory — offering distinct interaction styles tailored for onboarding, safety protocols, and complex regulatory workflows.
Living Textbooks from Regulations (Book)
Interactive document exploration: regulations are transformed into actionable learning material — not just something to read, but something to work with.
Personalised Learner Model
A three‑tier memory architecture (L1/L2/L3) with auditable traces — the system retains what the learner has already covered and adapts dialogue accordingly. This is not a "black box," but a transparent memory graph.
Enterprise Perimeter
- On‑premise deployment: Docker / Podman, rootless, single‑port exposure
- Local LLM support: Ollama, vLLM, LM Studio, NVIDIA NIM, and others
- Multi‑tenant workspace isolation
- REST / CLI / JSON API for seamless integration with corporate portals
- Channels: Telegram, Slack, Teams, and more via Partners
Conversational Tutoring Grounded in Your Documents
A unified agentic loop powered by RAG over your enterprise knowledge base: responses drawn exclusively from your internal regulations — never from the "Internet's memory." Supports PDF, DOCX, XLSX, PPTX; cross‑sectional relationships are mapped via GraphRAG / LightRAG.
Mastery Path Learning Trajectories
Step‑by‑step progression through topics with mastery gates: you move forward only after demonstrated understanding. This is not a linear "click‑through" course, but a governed, adaptive route.
Assessments & Question Bank
Automated generation and administration of questions with detailed explanations — designed for reinforcement and self‑checking.
Library of AI Mentors
Persistent mentors with predefined personas and memory — offering distinct interaction styles tailored for onboarding, safety protocols, and complex regulatory workflows.
Living Textbooks from Regulations (Book)
Interactive document exploration: regulations are transformed into actionable learning material — not just something to read, but something to work with.
Personalised Learner Model
A three‑tier memory architecture (L1/L2/L3) with auditable traces — the system retains what the learner has already covered and adapts dialogue accordingly. This is not a "black box," but a transparent memory graph.
Enterprise Perimeter
- On‑premise deployment: Docker / Podman, rootless, single‑port exposure
- Local LLM support: Ollama, vLLM, LM Studio, NVIDIA NIM, and others
- Multi‑tenant workspace isolation
- REST / CLI / JSON API for seamless integration with corporate portals
- Channels: Telegram, Slack, Teams, and more via Partners
A learning system with demonstrable mastery — engineered for industrial scale.
Target audience: corporate universities, large industrial holdings (including oil & gas and chemicals), and organisations where training functions as a risk-management instrument — not a checkbox in an LMS.
Core proposition: the evolution of Version 1.0 into a full‑fledged Intelligent Tutoring System (ITS), underpinned by a Hybrid AI‑Pedagogy Engine — a synthesis of the Soviet‑Russian pedagogical tradition and Western measurement‑driven algorithms. The system does not merely answer questions: it shapes actionable competence, verifies mastery through mathematical rigour, and provides HR with granular bottleneck analytics.
What 2.0 adds on top of 1.0
Target audience: corporate universities, large industrial holdings (including oil & gas and chemicals), and organisations where training functions as a risk-management instrument — not a checkbox in an LMS.
Core proposition: the evolution of Version 1.0 into a full‑fledged Intelligent Tutoring System (ITS), underpinned by a Hybrid AI‑Pedagogy Engine — a synthesis of the Soviet‑Russian pedagogical tradition and Western measurement‑driven algorithms. The system does not merely answer questions: it shapes actionable competence, verifies mastery through mathematical rigour, and provides HR with granular bottleneck analytics.
What 2.0 adds on top of 1.0
Capability | Business impact |
BKT — Bayesian Knowledge Tracing | Progression only when P(Know) > 0.95 — eliminates “clicked through and forgot” |
IRT + Zone of Proximal Development (ZPD) | Tasks calibrated to a workable difficulty corridor — Vygotsky’s philosophy becomes an algorithm |
Galperin’s stepwise formation of mental actions | Step-by-step skill formation: “walk me through the steps” + verification of the orienting basis of action — less illusory understanding on the shop floor |
Socratic mode + scaffolding | The tutor guides toward the answer instead of handing over a cheat sheet — develops thinking rather than replacing it |
Prerequisite graph | Non-linear trajectories: the system knows what must be mastered before a complex topic |
Dialogue-Aware RAG + knowledge graph | Answers linked to specific regulation paragraphs — auditable and verifiable |
HR Heatmap | Heat map of weak spots by document section — leaders see where learning breaks down |
Rights-Aware Guardrails + DLP | Deterministic protection: regulations and personally identifiable information (PII) do not leak into prompts unchecked |
Intent Router | Separates “emergency field query” from “training mode” — different response logic |
Digital learner profile | Neo4j + PostgreSQL + vector store — diagnosis and development in a single perimeter |
A containerised on‑premise solution (Docker) that automatically transforms PDF regulations and process flowcharts into interactive courses for your corporate LMS (WebTutor). The system deploys directly on the customer's servers, operates within an isolated perimeter with no cloud dependencies or external API calls: a local LLM (Qwen / Llama) extracts instructional semantics, while the compiler assembles a ready‑to‑import ZIP package containing tests, flashcards, and a metric‑reporting layer back to the LMS.
01
The Russian pedagogical tradition. Chatbook Tutor builds on methodologies that have shaped the country's educational system for decades: the Zone of Proximal Development (ZPD, Vygotsky), the Stage‑by‑Stage Formation of Mental Actions (Galperin), Developmental Education (Elkonin–Davydov), the Enlargement of Didactic Units (Erdniev), and Individual Educational Trajectories (IET). These foundational frameworks are embedded directly into the system's instructional logic — not as abstractions, but as executable pedagogical models.
02
A robust mathematical apparatus. Bloom's Taxonomy, Bayesian Knowledge Tracing (BKT), and Item Response Theory (IRT) are deployed to digitise, measure, and scale learning outcomes. In concert with the Russian pedagogical methodology, they function as the mathematical engine driving the entire system.
03
A Digital Learner Profile sits at the core of personalisation. The tutor ceases to be a mere "regulation search engine" and evolves into a true mentor — powered by four interwoven data layers: a mastery matrix, a cognitive‑behavioural profile, organisational context, and the full digital trace of dialogue history.
ADVANTAGES
Chatbook Hikma AI Lab
A Russian technology company and a full‑cycle, infrastructure‑grade AI factory. We convert fundamental R&D into trusted B2B and B2B2C products purpose‑built to advance the knowledge economy — ranging from interactive exploration of individual books to the design of personalised learning trajectories and the establishment of full‑scale corporate universities and knowledge systems.
Our core expertise lies in sovereign, secure, and culturally adaptive AI solutions for the digital transformation of publishing, education, and corporate learning — all powered by our proprietary Legal & Rights‑Aware RAG and Hybrid AI‑Pedagogy Engine. The product portfolio includes Chatbook Sales, Chatbook, and Chatbook Tutor.
The Russian Federation serves as both our home market and our primary R&D proving ground. The MENA region and partners across the Global South are our priority scaling destinations, driven by technology transfer and localised deployment.
Our core expertise lies in sovereign, secure, and culturally adaptive AI solutions for the digital transformation of publishing, education, and corporate learning — all powered by our proprietary Legal & Rights‑Aware RAG and Hybrid AI‑Pedagogy Engine. The product portfolio includes Chatbook Sales, Chatbook, and Chatbook Tutor.
The Russian Federation serves as both our home market and our primary R&D proving ground. The MENA region and partners across the Global South are our priority scaling destinations, driven by technology transfer and localised deployment.
LANA
MININA
MININA
CEO, co-founder
Mentor at the Skolkovo Foundation startup accelerator, international StartupBootcamp accelerator, certified tracker and ICF‑accredited coach. Holds multiple advanced degrees in pedagogy, marketing and advertising, and philology. Methodologist behind the Chatbook and Chatbook Tutor educational products, with over 20 years of experience in IT marketing and sales, including international markets.
MAKSIM
FEDYUKOV
FEDYUKOV
CTO, Co-founder
Mentor for startups at the Skolkovo Foundation, candidate of physical and mathematical sciences (PhD equivalent), author of registered patents in computer vision and machine learning, creator of educational courses, and expert in the deployment of generative and agentic AI. Over 20 years of experience in IT project management, including international markets.
EVGENIY
VERSHININ
VERSHININ
Software Architect
Mentor for startups at the Skolkovo Foundation, expert at the National Technology Initiative (NTI), Candidate of Technical Sciences (specialisation: Mathematical Modelling, Numerical Methods, and Software Suites), with over 20 years of experience at the intersection of mathematical modelling and commercial AI systems development.
TEAM
{CHATBOOK HIKMA AI LAB}
The knowledge economy of the future
The knowledge economy of the future
info@chatbook.pro
