When a team of insurance brokers receives more than 500 emails per day from clients, it quickly becomes difficult to keep things organized and make sense of it all. That’s where Libero comes in: a solution that summarizes and classifies all incoming emails. Beyond that, Libero also categorizes them properly within the client database (CRM). All of this, which brokers and their assistants used to do manually, is now fully automated, freeing up a tremendous amount of time every day. Through this presentation, we want to bring you into the heart of Libero’s design and the key decisions made during its development: its architecture, the challenges, the specific requirements of the insurance industry, the solution’s evolution, and more. And most importantly, we want to open up a discussion with you, the community, around this question: - How do we build and deploy AI solutions that can actually be maintained and evolve over time? This question is more important now than ever, with the rapid evolution of AI solutions, products and services that hit the market each week. The rhythm of innovation (and sometime fluff…) is astonishing! How do we continue building solutions that stay relevant and keep delivering business value? As a service provider, iuvo-ai is constantly balancing innovation with pragmatism. Every client has a different level of technical maturity, infrastructure, and internal talent. In that reality, the real challenge isn’t just getting a solution to work; it’s making sure it can live on. How do we design architectures that are flexible enough to evolve as the ecosystem changes, but simple enough for our clients to own and maintain? How do we make decisions that reduce friction when the next API version drops or when the internal IT team needs to take over? Those are the questions we wrestle with every day when bringing AI into production, and we’d love to exchange ideas and lessons learned with you 🙂