What do you think?
Agenci AI bazujący na modelach językowych. Istota, konfiguracje, zastosowania
Agenci AI to algorytmy wykorzystujące modele językowe jako reasoning engine. Są one zdolne do postrzegania otoczenia, rozumowania i podejmowania decyzji, co czyni je przydatnymi w wielu dziedzinach biznesu, między innymi:
- w spersonalizowanej obsłudze klienta,
- w automatyzacji procesów biznesowych,
- w zaawansowanej analityce biznesowej,
- we wspieraniu ludzi pracujących w takich działach jak HR czy R&D.
Użycie agentów AI może przynieść firmom wymierne oszczędności, usprawnić proces podejmowania decyzji i w efekcie zagwarantować trwałą przewagę konkurencyjną.
- w spersonalizowanej obsłudze klienta,
- w automatyzacji procesów biznesowych,
- w zaawansowanej analityce biznesowej,
- we wspieraniu ludzi pracujących w takich działach jak HR czy R&D.
Użycie agentów AI może przynieść firmom wymierne oszczędności, usprawnić proces podejmowania decyzji i w efekcie zagwarantować trwałą przewagę konkurencyjną.
144 pages, Paperback
Published January 1, 2025
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 1 of 1 review
January 11, 2026
What really hides behind the term “AI Agents” — hype, tools, or a new way of building systems?
Mariusz Hofman’s AI Agents tries to answer exactly that question — although it does so in a way full of contrasts.
First of all, it’s hard to tell who this book is for.
It’s not for engineers — there’s too little technical depth where it actually matters.
It’s not for business readers either — because suddenly you’re thrown into parameters, configurations, and code.
It lives somewhere in between: at the architecture level, but without a deep dive into real implementation.
And that is especially visible at the beginning.
On the one hand, you get a broad, high-level overview of AI agents: what they are, how they work, and how they can be combined into systems. That part is genuinely valuable — especially for someone who already knows LLMs but hasn’t yet thought about agents in a more systemic way.
On the other hand, you suddenly find yourself deep in code and configuration, sometimes explained down to the level of translating parameter names. That makes the book harder to read rather than more insightful.
So the book is full of contrasts.
At one moment it’s wide, conceptual, and engaging — and the next it feels like framework documentation.
The book improves as you go. The further I read, the more interesting it became. Unfortunately, the first chapter — the one visible in the online preview — sets the bar too low. From about the middle onward, it gets much better: more substance, a better balance between theory and practice, and more meaningful takeaways.
On the plus side, the book is surprisingly future-proof.
MCP, A2A, agent architectures — the author anticipated things that have already happened. In that sense, this is a book about a paradigm, not about a specific tool. That’s a real strength.
On the downside, there are noticeable editorial issues and inconsistencies in the references. The bibliography is hard to use, and given the number of examples, a proper code repository would have been very helpful.
In summary:
This is a good book — especially in the second half. It’s worth pushing through, even if the beginning doesn’t impress. But it should be read as an introduction to the world of AI agents, not as a self-contained handbook.
This is not a book that will “teach you how to build agents.”
It’s a book that helps you understand how they work — and gives you a starting point to begin building your own.
Mariusz Hofman’s AI Agents tries to answer exactly that question — although it does so in a way full of contrasts.
First of all, it’s hard to tell who this book is for.
It’s not for engineers — there’s too little technical depth where it actually matters.
It’s not for business readers either — because suddenly you’re thrown into parameters, configurations, and code.
It lives somewhere in between: at the architecture level, but without a deep dive into real implementation.
And that is especially visible at the beginning.
On the one hand, you get a broad, high-level overview of AI agents: what they are, how they work, and how they can be combined into systems. That part is genuinely valuable — especially for someone who already knows LLMs but hasn’t yet thought about agents in a more systemic way.
On the other hand, you suddenly find yourself deep in code and configuration, sometimes explained down to the level of translating parameter names. That makes the book harder to read rather than more insightful.
So the book is full of contrasts.
At one moment it’s wide, conceptual, and engaging — and the next it feels like framework documentation.
The book improves as you go. The further I read, the more interesting it became. Unfortunately, the first chapter — the one visible in the online preview — sets the bar too low. From about the middle onward, it gets much better: more substance, a better balance between theory and practice, and more meaningful takeaways.
On the plus side, the book is surprisingly future-proof.
MCP, A2A, agent architectures — the author anticipated things that have already happened. In that sense, this is a book about a paradigm, not about a specific tool. That’s a real strength.
On the downside, there are noticeable editorial issues and inconsistencies in the references. The bibliography is hard to use, and given the number of examples, a proper code repository would have been very helpful.
In summary:
This is a good book — especially in the second half. It’s worth pushing through, even if the beginning doesn’t impress. But it should be read as an introduction to the world of AI agents, not as a self-contained handbook.
This is not a book that will “teach you how to build agents.”
It’s a book that helps you understand how they work — and gives you a starting point to begin building your own.
Displaying 1 - 1 of 1 review