Tobias Goeschel at DDD Europe 2025
summary
Agents (conversational LLMs) enable rapid prototyping of domain logic by letting teams start an agent and interactively explore behavior when domain knowledge is incomplete, guiding discovery of missing rules and alternative workflows. Built prototypes by sending scenarios and observing reactions, and by composing handlers to send different command or event sequences to the model to reveal new use cases. Works well for exploration and design feedback but not yet production-grade or fully reliable.
AI-written summary · about
related talks
21mDesigning in the age of the 10x engineer
41mBuilding LLM Agents – Challenges and Learnings
49mAI Agents: Collaboration in Real-Life Applications
26mContext Engineering: Practical Techniques for Improving Agent Quality Today
45mAI agents in practice - beyond the prototype
23mLLM Agents in Production: Why Go is the Missing Piece
39mLLM Agents Gone Wild (And How to Tame Them with Quarkus)
35mBuilding LLM-based Agents