Data & IT Transformation Intern
6‑month internship within the Tech & Ops department of one of Europe’s largest insurers, focused on data‑driven IT performance, automation and governance of IT projects.
About the Role
Embedded in the MOQ team (“Methods, Tools & Quality”), I worked on the operational backbone of Generali’s IT project governance. My scope sat at the intersection of reporting, automation and AI tooling to help the organization scale its transformation method (REPERE) and secure the quality of data used for decision‑making.
Key Missions
Performance tooling for IT projects: built a catalog and prototype interface (Python + Streamlit) to map all Tech & Ops Power BI reports, making best practices discoverable and reusable across teams.
Automation at scale: identified manual, low‑value tasks and deployed ~15 Power Automate flows and PowerApps, saving an estimated ~10 man‑days per month for the department (relances, reporting, HR updates, PI Planning logistics, etc.).
REPERE training automation: redesigned end‑to‑end training operations (invitations, eligibility, reminders, attendance, satisfaction surveys, documents) via an ecosystem of flows and a PowerApps interface for trainers, drastically reducing operational friction and errors.
Contributed to the broader digital transformation strategy by identifying optimization levers through data-driven audits
KPI governance & data quality: co‑designed a new process for KPI collection and traceability (Excel + PowerApps + Power BI) with granular access rights, validation rules, and full history of changes — reducing manipulation risks and restoring trust in steering dashboards.
AI agents for IT governance: developed a conversational agent (Copilot Studio) for Tech & Ops project managers to query internal documentation and processes in natural language, centralizing fragmented knowledge and reducing repetitive questions.
Advanced experimentation: designed a robot using AI Builder to assess completeness of initiative sheets; the project surfaced tooling limits (licensing, accuracy ceiling) but clarified what would be needed for a robust, industrial solution.