How agentic AI is reshaping team structures: analyzing AWS talks and thousands of publications

How agentic AI changes team structures: analysis of AWS presentation and thousands of publications

The author of the Habr material, who previously deliberately avoided topics of frontier models, autonomous agents, and conversations about AI-native companies, explained that after Amazon Web Services (AWS) presentation about teams in the agentic AI world, he decided to analyze thousands of publications about how major technology companies envision the future structure of teams.

Why the author avoided this topic

The author deliberately avoided three topics for years: frontier models, autonomous agents for their own sake, and conversations about AI-native companies. The reason, according to him, is an excess of marketing and a lack of engineering specificity around these topics.

"I have long deliberately avoided three annoying topics: frontier models, autonomous agents for their own sake, and conversations about AI-native companies.

Because there is too much marketing around them and too little engineering specificity," writes the author of the article on Habr.

What changed the AWS presentation

The AWS presentation about teams in the world of agentic AI made the author reconsider his approach to the topic. According to him, the interesting question is not which AI model will ultimately win on the market, but which operational model of team organization will survive the cheapening of software development and whether the industry has already come up with something worthwhile on this subject.

• Occasion for the article — AWS presentation about teams in the agentic AI world (video on YouTube)
• Three topics that the author previously avoided — frontier models, autonomous agents for their own sake, conversations about AI-native companies
• Format of the material — analysis of thousands of publications about the future structure of teams
• Sources of publications — major technology companies, abbreviated by the author as MAANG (Meta, Amazon, Apple, Netflix, Google)

About the article

The material is an analysis of thousands of publications about what major technology companies, which the author refers to by the acronym MAANG (Meta, Amazon, Apple, Netflix, Google), think about the future structure of teams and organization of development in conditions of code cheapening thanks to AI agents. The central question of the article, according to the author's formulation, is not the choice of a winning model, but the search for an operational model of a team resistant to code writing automation.

The author emphasizes that he is interested not in "which model will win", but in which operational model of development will survive as code writing becomes cheaper — and whether the industry already has a working answer to this question. Such a formulation of the question, according to him, distinguishes the analysis of the AWS presentation from typical conversations about frontier models and AI-native companies, which he deliberately avoided before: whereas people usually argue about which model is smarter, the AWS presentation raises the question of the structure of the team itself and the distribution of roles within it.

What this means

The discussion is shifting from the question "which AI model is better" to the question "how to reorganize the team and development processes" when code writing ceases to be the main bottleneck — and it is precisely this, according to the author, that became the reason for him to analyze the position of major technology companies on this topic.