From Chaos to Clarity: The AI Teammate That Reads Your Code and Writes Your Tickets
Quickie (BEGINNER level)
MC 2
This talk introduces SMITH, an internal agentic AI assistant that reads your actual codebase and then transforms vague Jira tickets into fully refined, sprint-ready stories, with proper acceptance criteria, implementation guidance, and realistic estimates. Through a live demo, it explores how an autonomous agentic AI system can replace a 45-minute refinement meeting with a 30-second intelligent analysis, all running 100% locally with zero data leakage.
Agile teams worldwide lose countless hours in refinement ceremonies trying to clarify ambiguous requirements, yet the technical answers are often already embedded in the existing codebase. Recent advances in agentic AI, retrieval-augmented generation (RAG), and local LLM inference have made it possible to build intelligent assistants that autonomously reason over both project management artifacts and source code without sending a single byte to the cloud. This talk emerges from real-world experience building and using AI solutions across enterprise projects, where privacy constraints rule out cloud-based AI tools and where bridging the gap between "what to build" and "what already exists" has a direct, measurable impact on team velocity and ticket quality.
Agile teams worldwide lose countless hours in refinement ceremonies trying to clarify ambiguous requirements, yet the technical answers are often already embedded in the existing codebase. Recent advances in agentic AI, retrieval-augmented generation (RAG), and local LLM inference have made it possible to build intelligent assistants that autonomously reason over both project management artifacts and source code without sending a single byte to the cloud. This talk emerges from real-world experience building and using AI solutions across enterprise projects, where privacy constraints rule out cloud-based AI tools and where bridging the gap between "what to build" and "what already exists" has a direct, measurable impact on team velocity and ticket quality.
Konstantinos Elemenoglou
T-Digital by Deutsche Telekom
Konstantinos works for Deutsche Telekom as a Solution Designer, helping transform business needs into scalable and smarter digital solutions. Passionate about AI-driven innovation, intelligent automation and human-centric technology, he enjoys exploring how new tools can simplify systems, empower teams, and create better user experiences. Ambassador of “Let’s try this with AI” (coffee still runs on manual mode).