Yes — and that's the main thing that changed in 2025/2026. AI models (Claude, GPT) can now write functional code, integrate APIs, and design architecture. You need business understanding and the ability to manage an AI team — not necessarily programming experience.

AI models reached a point in late 2025 where they're not a "helper tool" but a full development partner. They can write entire systems, integrate data sources, test code — and do it in parallel, 24/7. What used to require a team and months is now possible in days.

Prototype — from a few hours. Dashboard with 6 data sources — in a day. Full CRM system — in a few weeks. The main time cost isn't writing code, but defining business logic and data migration.

Business understanding and a basic grasp of infrastructure — servers, databases, APIs, deployment. Programming is not required — AI does that. But you need to understand what happens "under the hood" to make the right decisions. And if something isn't clear — you can always ask AI.

An entire ecosystem that's typical of traditional development falls away — both outsourced and in-house. Project managers, business analysts, QA engineers, DevOps specialists, separate frontend and backend developers — AI takes over most of these roles. What stays with the human: business vision, architecture decisions, and quality control.