AI in practice: from a start-up built with Claude Code to an agent workflow from meeting to prototype

Welcome everyone! 👋 If you're looking for practical examples of how AI can actually help you build things and work smarter - this is the right place. 📌 What to expect: Practical AI use cases with live demosDirect Q&A with the speakers - ask anythingRelaxed coffee-break networking with people experimenting with AIConcrete inspiration you can start using the same day🍽 Lunch together: After the official part, join us for lunch and continue the conversations with the speakers and fellow participants 🎤 𝐓𝐚𝐥𝐤 𝟏: 𝐂𝐥𝐚𝐮𝐝𝐞 𝐓𝐨𝐥𝐝 𝐌𝐞 𝐈𝐭 𝐖𝐚𝐬 𝐌𝐲 𝐂𝐄𝐎: 𝐅𝐢𝐞𝐥𝐝 𝐍𝐨𝐭𝐞𝐬 𝐟𝐫𝐨𝐦 𝐅𝐨𝐮𝐫 𝐌𝐨𝐧𝐭𝐡𝐬 𝐨𝐟 𝐁𝐮𝐢𝐥𝐝𝐢𝐧𝐠 𝐚 𝐒𝐭𝐚𝐫𝐭-𝐮𝐩 𝐰𝐢𝐭𝐡 𝐂𝐥𝐚𝐮𝐝𝐞 𝐂𝐨𝐝𝐞 Four months ago, Alaa started building his AI start-up, Visioth, with Claude Code as his main collaborator: the website, company email, security, a cloud server and a small team of AI agents. It moved fast, it was sometimes confidently wrong, and one day Claude told him it was his CEO. Another day, his own account recovery codes ended up somewhere they should never have been. This talk distils those four months into the working habits he wishes he had known on day one, shown through live demos with almost no slides: why a single AI answer is only a first draft and how to make several agents check each other, how to give Claude a lasting memory of who you are and the rules you follow, how to stop burning through usage limits, when a small script beats the model, and the safety checks that catch what the AI misses. For founders, builders and anyone using AI for real work. No coding background needed. 🙋 Alaa Khashoun: mechatronics engineer (MSc, Warsaw University of Technology and FAU Erlangen) and founder of Visioth, a Warsaw start-up building quality checks for the data that trains computer-vision AI. Builds with Claude Code, AI agents and Python at visioth.com. 🎤 𝐓𝐚𝐥𝐤 𝟐: 𝐀𝐈 𝐀𝐠𝐞𝐧𝐭𝐬 𝐚𝐭 𝐖𝐨𝐫𝐤: 𝐅𝐫𝐨𝐦 𝐌𝐞𝐞𝐭𝐢𝐧𝐠 𝐭𝐨 𝐕𝐞𝐫𝐢𝐟𝐢𝐞𝐝 𝐎𝐮𝐭𝐜𝐨𝐦𝐞 (40-minute practical workshop for AI to Grow) Andrew Tomin will show how an AI agent finds decisions and gaps in a realistic meeting transcript, builds a working brief and creates a clickable prototype after human review. The session starts with a simple comparison between a browser-based AI chat and an agent that can work with goals, files, context and tools. Andrew then demonstrates one complete workflow:Meeting transcript → decisions and gaps → human clarification → working brief → Human Check → prototype → outcome validation The prototype is used to check whether the output solves the original task. The goal is not to generate more documents, but to show how a working brief, a Human Check and a prototype keep the result grounded in the source. Designed for product and project managers, marketers and content teams, operations managers and team leads, founders and independent experts. No coding experience required. Participants will receive: A reusable meeting-to-prototype workflowA working brief templateA prompt sequence for each stepA Human Check checklist 🙋‍♂️ Andrew Tomin: AI Product Manager, AI adoption consultant and founder of AISandboxLab.com, a hands-on AI bootcamp for non-coders. 14+ years building B2B products, including at EPAM, WorkFusion, Wargaming and Inspectorio. Helps teams adopt AI in day-to-day operations and move beyond browser-based AI chats, using Codex, Claude Code and Cursor to build workflows where agents work with project files and context and return results for human review. 💶 Entry fee: 30 zł (cash or card) 👉 Join the AItalks2Grow community to stay updated on upcoming events Open Google Form