Vienna Agentics Meetup: Building Agents That Work

Description

AI agents are easy to demo and hard to trust. They can act on stale context, hide what they are doing, and say “done” before the work is verified. On July 30, five practitioners from London, Seattle, Novi Sad, and Kansas City will bring their perspectives to Vienna. Their talks examine what agents need beyond the model: usable interfaces, capable people and organizations, reliable verification, organizational memory, and infrastructure grounded in real systems. Each session follows the same format: a 10-minute presentation followed by 5 minutes of questions. After the talks, we will continue with an hour of open networking and informal project sharing. The event is free, volunteer-run, and noncommercial. Schedule 18:00 — Doors open Arrive, order food or a drink, and meet other attendees. 18:30–20:00 — Lightning talks and Q&A 20:00–21:00 — Open networking and informal project sharing Continue the conversations, compare approaches, open a laptop, or show people something you are building, testing, or trying to understand. Bring something to share Bring a working system, prototype, useful failure, framework, or unresolved problem. After 20:00, attendees are welcome to share projects informally, ask for feedback, or find collaborators. You do not need a polished deck. Please share what you learned, not what you sell. Who should attend? This meetup is for people who build, deploy, evaluate, design, govern, research, or prepare organizations for agentic systems. You do not need to be an Agentics Foundation member or an expert. Join in Vienna or online In person: Pappala Pub, private event room U-Bahn-Bogen 156–157, 1090 Vienna Food and drinks will be available for purchase. Online: Register to receive the Zoom access link. Remote attendees can follow the programmed talks; networking and informal project sharing will primarily take place in the room. About the Agentics Foundation The Agentics Foundation is a global not-for-profit organization working to make AI innovation and education more accessible by supporting the development and deployment of open-source agentic AI systems. Its community brings together developers, researchers, founders, and practitioners through education, collaborative projects, events, and regional chapters. Talks Intro — Welcome to the Agentics Foundation A short introduction to the Foundation, its international community, and the role of open knowledge sharing, open collaboration, education, and open-source work in the development of agentic systems. Bence Csernak — From Screens to Relationships: Designing UX for Agents That Remember and Act(LinkedIn - GitHub) Traditional software waits for a user to click. Agents can remember context, initiate work, continue acting after an interaction, and make decisions on a user’s behalf. Bence will examine what this changes for product and UX design. Bence Csernak (ex EY, HSBC, NatWest) is an Applied AI Product Lead and Agentic AI Engineer with production experience designing and deploying LLM systems across healthcare, enterprise, and regulated environments. Member of the Board at Agentics Foundation (volunteer) Klara Hermesz — The AI Competency Framework & Diagnostic Tool for Enterprises and Individuals(LinkedIn - GitHub) Strategic AI upskilling needs a starting point and a clear map of where to go. Without a shared framework, individuals, teams, and organizations struggle to see where they stand, which capabilities their work requires, and what they should develop next. Klara will walk through the AI Competency Framework she developed with Adam Kovacs and demonstrate a diagnostic tool built around it for individuals and organizations. Klara Hermesz is co-founder and Chief Learning Strategist of the AI Enablement Academy. She previously built and scaled global learning programs at Meta and Uber, including Meta’s first metaverse education strategy for a 3,000-person global people organization and a redesign of Uber’s mentoring and coaching program for more than 50,000 employees. A self-confessed learning geek, she has built her career around a consistent conviction: the tools are rarely the hardest part; bringing people along is. Dragan “Profa” Spiridonov — Completion Theater: When Your Agent Says “Done” — and Lies(LinkedIn - GitHub) Every agent builder has watched an agent confidently claim success on broken work. This talk names the problem — completion theater — shows the failure modes behind it, and provides practical verification patterns to increase confidence in your agents’ results. Dragan “Profa” Spiridonov is Head of Agentic Quality Engineering at Cognitum One, creator of the PACTS and Agentic QE framework and platform. A 30-year IT veteran who builds quality into agentic systems rather than testing it afterward. Agentics Foundation Board Secretary and ambassador for the Serbian AF Chapter. Adam Kovacs — Surveillance or Observability? The Context Layer an AI Chief of Staff Needs(LinkedIn - GitHub) Organizations already generate messages, documents, tickets, code, meetings, and system logs. Those traces are not automatically usable knowledge. A chief-of-staff-style agent needs structured, current, permission-aware context. Adam will explain five properties that context layer needs: shared meaning, time, provenance, permission, and purpose. He will then use them to examine an uncomfortable question: when does useful organizational observability become workplace surveillance? The session is adapted from Adam and Klara’s 2026 Craft Conference presentation, “Get Your Systems to Define the Story of Your Organization: Is It Work Surveillance, or Hyper-Optimized Operational Excellence?” Adam Kovacs is co-founder and Chief AI Strategist of the AI Enablement Academy and a Co-founding Member and International Ambassador of the Agentics Foundation and its Seattle Chapter. He was a founding member of the AWS Talent Intelligence service and is Executive Director of OPEN Talent Society, a Seattle-based nonprofit exploring the future of work. His work focuses on enterprise AI adoption, agent-ready operations, organizational memory, and responsible governance. Chris Royse — Eyes, Hands, and Ground Truth: Building the Full Stack of Autonomous Agents — from the Database Up(LinkedIn - GitHub) Chris will trace an autonomous-agent stack from its data layer to computer use: how an agent represents associations, builds structured context around code and other domains, perceives software, and takes action through tools. Chris Royse is a systems engineer building the full stack of autonomous AI agents (20000 commits in the past year). He is the creator of Synapse, an open-code MCP server that gives any AI the eyes and hands it needs to use a computer like a human. Calyx is an association native database built using the mathematical formula for intelligence. My current work has been building projects on top of this intelligent database. Astrolabe applies Calyx to code across 158 coding languages. Poly tracks the flow of money across the globe and analyzes how it is all associated. Some of my other work with Calyx includes a Biomedical Hypothesis Engine as well as an engine designed for US legal court cases. Learn more about the Agentics Foundation