Many data professionals have experimented with AI coding assistants, only to find the results inconsistent. AI outputs seemingly relying more on "vibes" rather than solid engineering principles. To transform these tools from novelty chatbots into reliable workflow multipliers, developers must abandon simple prompt engineering in favor of systematic Context Engineering.

This session cuts through the noise of AI terminology to demonstrate a practical, battle tested, end-to-end Agentic Workflow. We will ignore the theoretical buzzwords and build a live, production-ready data platform from raw source to a Medallion architecture, finishing with a published Power BI dashboard. We will demonstrate exactly how to orchestrate AI agents, leveraging Model Context Protocol (MCP), skills, custom personas, and specific tool integrations to handle complex data pipelines.

You will leave with a concrete understanding of:

- Context Engineering: Mastering the shift from "asking" the AI to "architecting" the AI's environment for reproducible results.
- Configuring and customising agentic tools (like GitHub Copilot or Claude) with specific personas and skills to level up your existing work
- How to Deliver a data platform solution (Ingestion → Silver/Gold → Reporting) purely through AI augmented workflows.
- Identifying which parts of the "Agentic" landscape are hype and which are essential for accelerating outcomes.