COMMON POWERUp 2026

Artificial Intelligence, by nature, is probabilistic. Even in narrow tolerance bands, AI models can generate varying results for the same inputs — not due to bugs, but because of the statistical nature of learning. While this flexibility allows AI to deal with complex problems, it also introduces uncertainty — a dangerous liability in high-stakes business contexts. In using Project BOB or any LLM to modernize or transform code you can get mistakes, bugs or hallucinations in the new code, and sometimes those errors can be very tricky to be identified and corrected, it could be like looking for a needle in a haystack. In this session, we will see which methodologies to adopt to master these problems, to control any AI mistakes and also to influence and guide AI to fix and avoid them.

Learning Objectives:

For project managers and programmers this session will show what framework to settle and which methodologies to adopt in order to optimize quality outcomes and reliability from LLM, especially in a code-assistant and code-generator business perimeter.