From Pyramids to Diamonds: How AI reshapes the Organization of Software and IT Consulting companies (and not only…)

The traditional organizational model in software development and IT consulting has long resembled a pyramid: a small layer of decision-makers and architects at the top, supported by middle management, and a broad base of developers executing the work.

This structure worked well in a world where value is created primarily through manual coding and implementation. But with the rise of large language models (LLMs) and AI agents capable of reasoning, generating code, and automating tasks, the game is rapidly changing.

We are now witnessing the emergence of a new model that resembles a diamond more closely. In this paradigm, the broadest layer of the organization is no longer at the bottom but in the middle. The core value creators are no longer just software developers but domain experts, knowledge workers, and AI engineers who translate complex expertise into structured prompts, tasks, and workflows for AI agents.

Knowledge and Natural Language become the New Code

In the diamond model, the organization's foundation is deep domain knowledge: industry-specific insights, business logic, regulatory nuances, and strategic context. This knowledge is no longer "thrown over the wall" to developers. Instead, it's actively curated, structured, and encoded into prompts and data that can guide LLMs and AI agents.

The traditional developer's role is evolving into something new: AI/LLM Engineers, who act as translators between human expertise and machine reasoning. They are responsible for transforming domain knowledge into instructions that LLM-powered systems can understand and execute effectively. They don't just write code, they design prompts, orchestrate agent workflows, and fine-tune outputs based on real-world feedback.

The Diamond Structure: A New Division of Labor

In this new structure, the layers of the organization shift:

• Top: Strategic leadership and governance remain essential but now include broad and real knowledge of AI
• Middle (Broadest Layer): A large, empowered group of domain experts and AI engineers who create, manage, and optimize knowledge flows. This is where innovation and value creation happen.
• Bottom: A way smaller team for Traditional Coders and IT Operations.

This model is not just about efficiency, it's about scaling intelligence.

The impact on the Cybersecurity and IAM industries

AI will reshape our beloved Cyber and IAM industries as well.

• Software vendors: Hire people who know your domain and let them write very good documents, as Natural Language is the new code.
• Consulting and system integrators: forget about hiring “juniors” and achieve your EBIT thanks to their endless dedication to the company. SOC L1 and L2? Gone. Junior Building connectors in IAM platforms? Gone.
• Auditors, Risk & Compliance Professionals: Get ready to Audit AI-based applications. It's better than 'traditionally software-coded' applications, believe me. Policies are in Natural Language, so get prepared to write clear documents that can be reasoned upon by Large Language Models.

Conclusion

The transition from a pyramid to a diamond organization is more than a reorganization: it's a fundamental rethinking of how value is created in the age of AI. As knowledge and natural language become the new code, and AI agents take on increasing responsibility for reasoning and execution, organizations must embrace this shift or risk being left behind.