
Your AI Transformation Has Too Many People
AI is not just making software teams faster. It is removing the need for much of the team. What I built, what happened, and exactly where the evidence stops.
By letting the person who understands the problem build the solution directly.
Of ~800 employees
The company’s operating layer is now written by the people who operate it.
Arguments about how organisations change when the people who understand a problem can build the software that solves it.

AI is not just making software teams faster. It is removing the need for much of the team. What I built, what happened, and exactly where the evidence stops.
What changes when the people who understand a problem can build for it themselves — the numbers, the failures, and the questions still open.

Vienna AI Engineering — how I built an internal platform without being a developer, the overnight agent loop, and the bug that took ten failed fixes before a different model found it.
Daniil Aizenshtat leads Strategy & Execution at EXANTE, an international regulated broker of roughly 800 people. He built the company's internal application platform and led its adoption across the business, where colleagues now build and run the software their own work depends on.