[{"title":"Executive Semantic Model Factory with Conversational BI","description":"Every business unit builds its own Power BI reports with inconsistent metrics and definitions, so executives cannot get one version of the truth and non-technical users cannot self-serve. Participants build a semantic model factory on Fabric: consolidate redundant reports, harmonize KPIs into a Business Questions Library, generate and validate reusable semantic models with tested DAX, derive a Fabric IQ ontology capturing entities and relationships, and layer a conversational interface offering natural-language-to-query plus GenAI dashboard summaries ready to paste into a business email. Participants will deliver a consolidated KPI dictionary, one certified semantic model with validated DAX measures, a published ontology, and a Data Agent answering at least 10 executive questions correctly.","image_url":"https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcRftMJjUVXRQye64Xwjg8-lEw0pAgHeicwbiz6SrDvQyVgli1qc6A-cFXc&s=10","hackathon_type":"htb","challenge_scenario":"Group CFO, Woodgrove Bank: 'I asked four business units for net revenue last Tuesday and got four numbers and four explanations of why the other three were wrong. My board pack takes nine days to assemble and I still cannot ask it a follow-up question.' Each business unit maintains its own Power BI estate with locally defined metrics, so definitions drift, reports duplicate, and the analytics team spends its week reconciling instead of analysing. Non-technical executives have no way to self-serve, which pushes every ad-hoc question back into the same queue.","learning_objectives":null,"expected_outcome":null,"industry":"Cross-Industry (Enterprise Analytics and Executive Reporting)","duration":"6 Hours · Teams of 4-5 (mixed BI developer / analytics engineer / business analyst)","duration_days":1,"duration_hours":"6.00","level":"intermediate","hands_on_lab":true,"validation":null,"intended_audience":null,"entry_requirements":null,"solution":null,"cloud_technology_stack":"Microsoft Fabric, Fabric Warehouse, Fabric Lakehouse, OneLake, OneLake catalog, Power BI semantic models (Direct Lake), Power BI Copilot, Fabric IQ (ontology, NL2Ontology), Fabric Data Agent, Microsoft 365 Copilot, Microsoft Purview (endorsement and certification, Unified Catalog), Fabric Git integration","tags":"Fabric, Fabric IQ, Power BI, M365 Copilot, Purview, Semantic Model","frontier_core":"core","solution_area":"Cloud & AI Platforms","conversation":"• Enterprise & SMC: Build a unified, governed data and AI estate | • SMB: Modernize your data to drive results and leverage AI | Solution Play: Unify Your Data Platform","conversation_number":3,"secondary_conversations":null,"format_long":null,"scenario_company":null,"lab_structure_summary":null,"level_unlock_thresholds":null,"assessment":null,"languages":"English","rooms_link":null,"lab_guide_link":null,"readiness_type":"pr","source_document":null,"status":"active","card_icons":[],"id":"H66","created_at":null,"updated_at":"2026-08-26","track_units":[],"global_track_id":"a29d4a83-01c9-40a0-a71a-121ad2ad85b0","origin_system":"import_xlsx","origin_ref":null,"revision":null,"lab_status":"none","lab_provider":null,"lab_ref":null,"lab_built_at":null,"summary":null,"sync_state":"synced"},{"title":"Enterprise Knowledge Intelligence Platform & Service Assistant (RAG)","description":"Organizational knowledge is trapped in documents, SharePoint, ticketing systems and operational data, so field engineers and support staff waste hours retrieving procedures and precedent cases. Participants build a Fabric knowledge platform that combines structured and unstructured content: ingestion and batch summarization of tickets and documents into a Lakehouse, vector embedding and similarity search, and a customized RAG assistant that returns step-by-step guidance with citations at sub-second retrieval — with Graph RAG over an entity and knowledge graph as the stretch goal. Participants will deliver an indexed corpus of 1,000+ documents and tickets, a working retrieval pipeline with grounded citations, and an evaluated assistant scored against a 20-question benchmark.","image_url":"https://www.softwebsolutions.com/wp-content/uploads/2024/11/Agentic-rag-blog-image.webp","hackathon_type":"htb","challenge_scenario":"VP Service Operations, Tailspin Toys: 'My field engineers spend the first forty minutes of every callout reading. The procedure is in SharePoint, the last three times we fixed this machine are in the ticketing system, and the part history is in the ERP — and none of them talk to each other.' Institutional knowledge sits in documents, an ageing intranet, ticket free-text and operational telemetry, and the engineers who know where things are are the ones retiring. Retrieval time, not repair time, is the constraint on first-time-fix rate and on customer satisfaction.","learning_objectives":null,"expected_outcome":null,"industry":"Cross-Industry (Field Service, Support Operations and Knowledge Management)","duration":"6 Hours · Teams of 4-5 (mixed data engineer / AI engineer / service-operations SME)","duration_days":1,"duration_hours":"6.00","level":"advanced","hands_on_lab":true,"validation":null,"intended_audience":null,"entry_requirements":null,"solution":null,"cloud_technology_stack":"Microsoft Fabric, Fabric Lakehouse, OneLake, Fabric Data Factory, Fabric Notebooks (PySpark), Fabric AI functions (ai.summarize, ai.embed, ai.similarity, ai.extract), Fabric Data Agent, Fabric IQ (ontology for Graph RAG), Fabric Eventstream, Microsoft Foundry (models and evaluation), Microsoft 365 Copilot / Copilot Studio, Microsoft Purview, Power BI","tags":"Fabric, Fabric Data Agent, Fabric AI functions, Foundry, RAG, Graph RAG, Purview, Power BI","frontier_core":"frontier","solution_area":"Cloud & AI Platforms","conversation":"• Enterprise & SMC: Build a unified, governed data and AI estate | • SMB: Modernize your data to drive results and leverage AI | Solution Play: Innovate with Azure AI Platform & Apps","conversation_number":3,"secondary_conversations":null,"format_long":null,"scenario_company":null,"lab_structure_summary":null,"level_unlock_thresholds":null,"assessment":null,"languages":"English","rooms_link":null,"lab_guide_link":null,"readiness_type":"pr","source_document":null,"status":"active","card_icons":[],"id":"H67","created_at":null,"updated_at":"2026-08-26","track_units":[],"global_track_id":"45a70dd5-4289-4176-9c31-5bd7e2eca8c4","origin_system":"import_xlsx","origin_ref":null,"revision":null,"lab_status":"none","lab_provider":null,"lab_ref":null,"lab_built_at":null,"summary":null,"sync_state":"synced"},{"title":"GenAI Data Harmonization & Quality Guardrails","description":"Master data — vendor codes, commodity codes, SKUs and product descriptions — is inconsistent across 50K+ records and multiple ERPs, reporting depends on manual Excel reconciliation, and sensitive data sits unclassified, which blocks enterprise AI adoption outright. Participants build a Fabric-native harmonization service: zero-shot LLM classification with confidence scoring and human-in-the-loop review, automated data-quality rules and anomaly checks embedded in pipelines, sensitivity labelling and DSPM for AI posture in Microsoft Purview, and continuous model improvement from reviewer decisions. Participants will deliver a golden-record Gold table for at least 50,000 harmonized records, a confidence-banded review queue, a Purview data-quality scorecard with at least five active rules, and a certified semantic model fed only by governed data.","image_url":"https://www.genzeon.com/wp-content/uploads/2025/10/1wwF7P_3l9L1CAoPczf-sTg.webp","hackathon_type":"htb","challenge_scenario":"Chief Procurement Officer, Fabrikam Industrial Supply: 'Three ERPs, four ways of spelling the same supplier, and a spend report that two analysts rebuild in Excel every month and still disagree on. I have been told we cannot switch on Copilot until somebody can say which of these fields is confidential — and nobody can.' Roughly 50,000 vendor, commodity and SKU records are inconsistent across systems, category spend cannot be rolled up reliably, and sensitive commercial and personal data sits unclassified across the estate, so every enterprise AI proposal stalls at the security review.","learning_objectives":null,"expected_outcome":null,"industry":"Cross-Industry (Master Data Management, Data Governance and Compliance)","duration":"6 Hours · Teams of 4-5 (mixed data engineer / governance specialist / MDM analyst)","duration_days":1,"duration_hours":"6.00","level":"advanced","hands_on_lab":true,"validation":null,"intended_audience":null,"entry_requirements":null,"solution":null,"cloud_technology_stack":"Microsoft Fabric, Fabric Lakehouse, Fabric Data Factory (pipelines, Dataflows Gen2), Fabric Notebooks, Fabric AI functions (ai.classify, ai.similarity, ai.extract, ai.embed), Fabric Data Science / ML (MLflow experiments and model registry), OneLake, Microsoft Purview (Unified Catalog, data quality rules and scans, sensitivity labels, DLP, Data Security Posture Management for AI), Fabric Activator (Data Activator), Power BI","tags":"Fabric, Purview, Fabric AI functions, Fabric ML, Power BI, Data Activator","frontier_core":"core","solution_area":"Security","conversation":"• Enterprise & SMC: Build a unified, governed data and AI estate | • SMB: Modernize your data to drive results and leverage AI | Solution Play: Unify Your Data Platform","conversation_number":4,"secondary_conversations":null,"format_long":null,"scenario_company":null,"lab_structure_summary":null,"level_unlock_thresholds":null,"assessment":null,"languages":"English","rooms_link":null,"lab_guide_link":null,"readiness_type":"pr","source_document":null,"status":"active","card_icons":[],"id":"H65","created_at":null,"updated_at":"2026-08-26","track_units":[],"global_track_id":"d2d29684-c914-48ba-b3b4-4e91c53fda2e","origin_system":"import_xlsx","origin_ref":null,"revision":null,"lab_status":"none","lab_provider":null,"lab_ref":null,"lab_built_at":null,"summary":null,"sync_state":"synced"},{"title":"Multi-Agent Operations Analyst (Inventory / E&O)","description":"Excess and obsolescence analysis is monthly, manual and fragmented across systems with inconsistent product and SKU structures, so portfolio decisions lag by weeks and working capital stays locked up. Participants build an agent-swarm system on Fabric: a data-retrieval agent, an NLP-based SKU-matching agent that reconciles inconsistent descriptions, a calculation agent that moves E&O from monthly to daily, and a recommendation agent that produces actionable disposition options with tailored weekly reports per user group — a direct hit on the multi-agent and MCP judging criteria. Participants will deliver four cooperating agents with a documented orchestration contract, a daily-refreshed E&O position, and ranked disposition recommendations with quantified working-capital impact.","image_url":"https://www.shutterstock.com/image-photo/corporate-organisation-utilizes-multi-agent-260nw-2722398369.jpg","hackathon_type":"htb","challenge_scenario":"VP Supply Chain, Northwind Traders Manufacturing: 'Our excess and obsolescence number arrives on the eleventh working day, describes the world as it was five weeks ago, and takes two analysts a fortnight to build. By the time we decide to discount or scrap, the write-off has already happened.' Inventory sits across several systems with inconsistent product hierarchies and free-text SKU descriptions that defeat exact matching, so the same part appears three times and none of the three positions is right. Portfolio decisions lag by weeks and working capital stays locked in stock nobody will sell.","learning_objectives":null,"expected_outcome":null,"industry":"Manufacturing / Supply Chain (Inventory and Portfolio Analytics)","duration":"6 Hours · Teams of 4-5 (mixed AI engineer / data engineer / supply-chain SME)","duration_days":1,"duration_hours":"6.00","level":"advanced","hands_on_lab":true,"validation":null,"intended_audience":null,"entry_requirements":null,"solution":null,"cloud_technology_stack":"Microsoft Fabric, Fabric Data Agent, Fabric Lakehouse, OneLake, Fabric Notebooks, Fabric Data Science / ML (MLflow), Fabric AI functions (ai.similarity, ai.classify), Fabric Activator (Data Activator), Microsoft Copilot Studio (multi-agent orchestration, connected agents, MCP), Microsoft Foundry Agent Service, Agent 365 with Microsoft Entra Agent ID, Power BI, Microsoft Teams","tags":"Fabric, Fabric Data Agent, Copilot Studio, MCP, Agent 365, Fabric ML, Power BI, Data Activator","frontier_core":"frontier","solution_area":"Cloud & AI Platforms","conversation":"• Enterprise & SMC: Build a unified, governed data and AI estate | • SMB: Modernize your data to drive results and leverage AI | Solution Play: Innovate with Azure AI Platform & Apps","conversation_number":6,"secondary_conversations":null,"format_long":null,"scenario_company":null,"lab_structure_summary":null,"level_unlock_thresholds":null,"assessment":null,"languages":"English","rooms_link":null,"lab_guide_link":null,"readiness_type":"pr","source_document":null,"status":"active","card_icons":[],"id":"H68","created_at":null,"updated_at":"2026-08-26","track_units":[],"global_track_id":"3016bf2a-025b-4420-a7ee-78ae631e329c","origin_system":"import_xlsx","origin_ref":null,"revision":null,"lab_status":"none","lab_provider":null,"lab_ref":null,"lab_built_at":null,"summary":null,"sync_state":"synced"},{"title":"AI-Accelerated Legacy-to-Fabric Modernization Factory","description":"Enterprise data is spread across Snowflake, SQL Server, Databricks, SAP, legacy BI and flat files, and lift-and-shift migrations simply replicate the technical debt — ungoverned logic, opaque lineage and escalating cost — while sunset platforms create compliance and infosec exposure. Participants build a Fabric \"modernization factory\" in which LLM agents read legacy code and BI artefacts, emit human-readable business-logic and dependency documentation, auto-generate Fabric pipelines into a governed Medallion Lakehouse, and prove equivalence with automated reconciliation. Participants will deliver a documented dependency map for a legacy workload, at least 10 auto-generated and executed Fabric assets (pipelines, notebooks, Delta tables) across Bronze/Silver/Gold, and a reconciliation report evidencing row-count and measure parity against the legacy source.","image_url":"https://d274cmdd0goq94.cloudfront.net/wp-content/uploads/2026/08/AI-Driven-Application-Modernization_-A-Guide-to-Accelerate-Business-Transformation.webp","hackathon_type":"htb","challenge_scenario":"Chief Data Officer, Contoso Manufacturing Group: 'We have four analytics platforms, three of them are being sunset by their vendors, and nobody left in the building can tell me what the pricing logic in that 2011 stored procedure actually does. Every migration quote I get is eighteen months and a promise to move the mess somewhere more expensive.' The estate spans Snowflake, SQL Server, Databricks, SAP extracts, a legacy BI server and thousands of flat files. Lift-and-shift replicates ungoverned logic and opaque lineage into the new platform, cost keeps escalating, and the sunset platforms are now an audit and infosec finding rather than a roadmap item.","learning_objectives":null,"expected_outcome":null,"industry":"Cross-Industry (Data Estate Modernization and Migration)","duration":"6 Hours · Teams of 4-5 (mixed data engineer / migration architect / pre-sales lead)","duration_days":1,"duration_hours":"6.00","level":"advanced","hands_on_lab":true,"validation":null,"intended_audience":null,"entry_requirements":null,"solution":null,"cloud_technology_stack":"Microsoft Fabric, Fabric Lakehouse (Medallion Bronze/Silver/Gold), Fabric Data Factory (pipelines, copy job, mirroring), Fabric Notebooks (PySpark), OneLake, Fabric Warehouse, Migration Assistant for Fabric Data Warehouse, Fabric Copilot, Fabric AI functions, Power BI, Microsoft Purview (Unified Catalog, lineage), Azure DevOps / GitHub with Fabric Git integration","tags":"Fabric, Fabric Data Factory, OneLake, Purview, Power BI, Copilot, Migration","frontier_core":"core","solution_area":"Cloud & AI Platforms","conversation":"• Enterprise & SMC: Build a unified, governed data and AI estate | • SMB: Modernize your data to drive results and leverage AI | Solution Play: Unify Your Data Platform","conversation_number":7,"secondary_conversations":null,"format_long":null,"scenario_company":null,"lab_structure_summary":null,"level_unlock_thresholds":null,"assessment":null,"languages":"English","rooms_link":null,"lab_guide_link":null,"readiness_type":"pr","source_document":null,"status":"active","card_icons":[],"id":"H64","created_at":null,"updated_at":"2026-08-26","track_units":[],"global_track_id":"c1f92c77-065d-45a6-985f-42547e8a54f2","origin_system":"import_xlsx","origin_ref":null,"revision":null,"lab_status":"none","lab_provider":null,"lab_ref":null,"lab_built_at":null,"summary":null,"sync_state":"synced"},{"title":"Customer Intelligence Engine: Real-Time 360, Segmentation & Proactive Churn","description":"Companies are drowning in customer data but starving for insight — fragmented touchpoints, siloed CDP and analytics stacks, and churn signals discovered only after customers have already left. Participants build a Fabric-native customer intelligence engine: a Customer 360 in a Lakehouse spanning CRM, transactions and digital interactions, a customer ontology unifying entities and relationships across sources, dynamic segmentation and journey analytics, an agentic churn early-warning workflow that analyses longitudinal usage anomalies and raises automated alerts, and real-time activation hooks. Participants will deliver a resolved Customer 360 table, a published ontology, a scored churn model with explainability, and a Power BI dashboard flagging at-risk accounts three months ahead.","image_url":"https://www.cxtoday.com/wp-content/uploads/2026/05/customer-intelligence-platforms-behavioural-analytics-CX-real-time-customer-prediction-AI-driven-customer-insights-cx-today-2026-ai.jpg","hackathon_type":"htb","challenge_scenario":"Chief Customer Officer, Relecloud Subscription Services: 'We have a CDP, a warehouse, a marketing automation platform and a CRM, and I still find out a customer is leaving when the cancellation email arrives. We are paying four vendors to tell us the past.' Customer touchpoints are fragmented across systems that never agreed on an identity, segmentation is a quarterly spreadsheet exercise that is stale before it is circulated, and retention teams work from lists rather than signals. Churn is diagnosed after the fact, so every intervention is an apology rather than a save.","learning_objectives":null,"expected_outcome":null,"industry":"Cross-Industry (B2C and B2B Customer Analytics, Retention and Marketing)","duration":"6 Hours · Teams of 4-5 (mixed data scientist / data engineer / CRM or marketing SME)","duration_days":1,"duration_hours":"6.00","level":"advanced","hands_on_lab":true,"validation":null,"intended_audience":null,"entry_requirements":null,"solution":null,"cloud_technology_stack":"Microsoft Fabric, Fabric Lakehouse, OneLake, Fabric Eventstream, Fabric Real-Time Intelligence (Eventhouse, KQL), Fabric Activator (Data Activator), Fabric Data Science / ML (MLflow experiments, model registry), Fabric IQ (customer ontology), Fabric Data Agent, Power BI (Direct Lake), Microsoft Purview, Microsoft Copilot Studio","tags":"Fabric, Fabric IQ, Real-Time Intelligence, Eventstream, Data Activator, Fabric ML, Power BI, Purview, Copilot Studio","frontier_core":"frontier","solution_area":"Cloud & AI Platforms","conversation":"• Enterprise & SMC: Build a unified, governed data and AI estate | • SMB: Modernize your data to drive results and leverage AI | Solution Play: Unify Your Data Platform","conversation_number":8,"secondary_conversations":null,"format_long":null,"scenario_company":null,"lab_structure_summary":null,"level_unlock_thresholds":null,"assessment":null,"languages":"English","rooms_link":null,"lab_guide_link":null,"readiness_type":"pr","source_document":null,"status":"active","card_icons":[],"id":"H69","created_at":null,"updated_at":"2026-08-26","track_units":[],"global_track_id":"df9f3ab3-6f0d-4c10-a77d-0dedaccb74b9","origin_system":"import_xlsx","origin_ref":null,"revision":null,"lab_status":"none","lab_provider":null,"lab_ref":null,"lab_built_at":null,"summary":null,"sync_state":"synced"}]