Building with AI: three perspectives from Genos AI Meetup #2

Genos AI Meetup #2 just wrapped up, and this time we had three of our colleagues sharing what they’ve been working on with AI.

Andreea, Senior Business Application Developer at Global Mediator, showed us a live AI integration in Business Central. Using the Anthropic API, she built a system that reads PDF and XML documents attached to invoices and automatically extracts the relevant information, taking manual processing out of the equation. She also demonstrated how MCP (Model Context Protocol) can connect Business Central capabilities directly with AI models. A great example of something practical that can already be used in real-world projects.
Vadym, Senior Business Application Engineer at Global Mediator gave us a look at what’s possible with a multi-agent orchestrator using layered instructions, project-specific skills and external tools such as CI/CD and Jira through MCP. He also showed how persona-based agents can mirror different roles within a team. The bigger question behind his session was an interesting one: when can we actually trust AI to reliably orchestrate other AI agents?
Alexandru, .NET Developer at Zensai, took us through the different ways AI can support his daily work, from simple chat interactions to agents and fully autonomous workflows. He shared how he approaches building agent systems that can delegate tasks, work together and make smart use of tokens. A very practical look at how AI is becoming part of everyday development.

Three sessions, three different perspectives, and one clear takeaway: AI is becoming much more than a productivity tool. It is increasingly becoming part of the way we build, ship and review software.
And it’s great to see our own colleagues exploring what that looks like in practice.