04/10/2026
The WebExpo 2026 recap series continues with Lucerna Cinema. From the Great Hall's case for human judgement, this room got operational. 🏗️ Its speakers covered working systems: containers, MCP servers, product maps, typography, authorisation, research engines, Git, contract testing, funnels, refactoring, search visibility and AI adoption.
The common thread was clarity. Not tidiness for its own sake. Clarity as a standard operating procedure.
Start with the interfaces between systems. Cedric Clyburn showed how local containers can lead towards Kubernetes without making production feel like another planet. Ladislav Prskavec built an MCP server in Go and made authentication and prompt injection impossible to ignore. Sohan Maheshwar asked what an AI agent should be allowed to access 🔐, while Robin Pokorný demonstrated how contract tests catch disagreements between services before users do. Lucky Nkosi took the idea of an interface somewhere less expected: he used voice, gestures and bananas 🍌 to fly a drone with web APIs.
And interfaces only work if people can make sense of them. Oliver Schöndorfer looked at how typography ✍ affects reading and comprehension. Irena Zatloukalová argued for explicit expectations instead of leaving colleagues to decode unwritten rules. And Sergès Goma made Git history a team concern: the next developer needs to understand why something changed, not just see that it did.
Clarity also changes the decisions made before a product ships. Christiane Moser asked teams to pin down their users, needs and metrics before reacting to the next stakeholder request. Julian Della Mattia built a research engine to bring evidence into those decisions, and Katherine Corneilson offered a way through messy qualitative data 📊. Vladana Bačová looked for the leaks after sign-up rather than assuming more acquisition would fix the funnel. When Jan Toman moved a product from Flutter to React, the harder question was which parts deserved the most care.
AI ran through the room 🦾, but nobody got to skip the groundwork. Daria Rudnik asked what organisations are getting back from their AI initiatives and which decisions should remain human-owned. Aneta Holá and Aleš Moravec looked at search when AI assembles answers instead of listing links: will it understand and represent your brand correctly? Tomáš Braverman and Romana Trusinová tackled AI adoption in companies: build on existing strengths, decide what to outsource and help staff learn to use it.
Different talks, same useful habit: make the assumptions visible before someone else has to work around them.
