Job Description Most business users still navigate dashboards using slicers and dropdowns, which is efficient for analysts, but can be frustrating for business decision-makers. While AI features like Q&A and Copilot exist in Power BI, they are mainly design-time helpers and rarely used for real interactive exploration. This internship aims to design a search-first dashboard experience where a user can type a query like “Top 5 customers in Belgium” and instantly see the relevant visual, using existing Power BI capabilities (text filter, Q&A, bookmarks, dynamic titles, smart narratives). But, the project must also explore the use of Model Context Protocols (MCPs) to describe and expose the underlying Power BI data model in a structured, machine-readable way. This enables more intelligent, context-aware search and interaction, bridging traditional BI with conversational and AI-driven analytics. Think “Copilot-like experience,” but actually working today. Objective By the end of the internship, the student will: Build a functional search-driven Power BI dashboard prototype Combine Q&A, text filtering, bookmarks and dynamic visuals Conduct usability testing comparing slicers vs search-first interaction Prototype a lightweight MCP-like metadata layer describing the Power BI model Experiment with using MCPs to guide search intent parsing (linking queries like ‘top customers in Belgium’ to relevant DAX measures and visuals) Deliver a reusable framework internal demo for Sopra Steria consultants