Artificial intelligence has generated more noise than clarity. Some call it just another bubble; others are afraid of losing their jobs; the largest consulting firms are selling it as the savior of business in an age of crises.
This book gives you the real picture — from a manager's chair, not a lab. What AI is and how it learns. What it can and cannot do. The technical, organizational and psychological problems that come with it, and why up to 90% of AI products end up unwanted. And what to actually expect over the next 10–20 years.
Inside the Book
1. What AI Actually Is — and What It Cannot Do
The groundwork, in plain language: what artificial intelligence is, how models are trained, and the flaws every current solution shares. Then the map — AI sorted by the tasks it solves and by its "strength" and ability to handle uncertainty, along with copilots, AI agents and multi-agent systems. Finally, what narrow AI is genuinely good at today and where the trends are heading, so your expectations match reality. Distorted expectations are the number one cause of failed digital projects.
2. Generative AI and Large Language Models
What generative AI really is, what goes wrong with it, and which built-in limitations cause those problems — hallucinations included. Then LLMs specifically: why they took off, how they work under the hood, what RAG and fine-tuning are for and when each is the right answer, and the practical rules for writing requests that get you a usable result.
3. Regulation and Safety
Why AI development worries people, what regulators have actually done so far, and where the rules are heading. The safety half is written as a working method: the unacceptable events you must rule out, the scenarios that lead to them, the underlying causes — and the crash tests and requirements to put on any AI solution before you let it near your business.
4. Digital Advisors and What Comes Next
For business, the main value of AI is prediction, recommendation and decision support. This chapter covers decision support systems and digital advisors: how they are built, the technologies behind them, why so many of them disappoint, and what to do about it. Then a look further out — neuromorphic and quantum AI, and what they would actually change.
5. AI Inside the Management System
The heart of the book, and it works in both directions: how AI changes your strategy, organizational structure, business processes, project and product management, internal communication, data and cybersecurity, regular management practices, lean and the theory of constraints — and what each of those disciplines demands of AI in return. Includes the seven deadly sins of digitalization: the mistakes behind 90% of failed projects, starting with implementing a technology because it is fashionable rather than because someone defined the benefit.
6. Function by Function, Company by Company
Where AI actually lands: digital business and marketing, production, analytics, digital products, new ways of working. Then real cases at every scale — AI inside 1C solutions, marketplaces of ready-made AI tools, digital advisors for project management, generative AI for content creators and custom assistants for small business; and for larger companies, big data with IoT and AI on China's railways, applications in power supply, at Nornickel and EuroChem, in ore processing, technical support, food service and oil project planning. Each with recommendations you can act on.
7. How to Run an AI Project — and the Roadmap
The delivery mechanics: the project lifecycle, defining goals and the effects you expect, the roles you need, and three separate algorithms depending on scale — a local departmental project, a cross-functional one, and a strategic one. It closes with a systems approach to adoption, a roadmap linking AI to your other initiatives, and an honest look at the hype: we wanted robots to do the dishes so we could create, and instead the robots are creating. Which jobs are really at risk, and whether you are ready for what that changes.