A major manufacturer needed to establish a sustainable data modeling practice within Fabric to deliver real-time, self-service business intelligence to field sales teams via conversational AI. Using a coaching-focused approach, Trility establishes a reliable data modeling pattern that enables the client to arm the sales force with critical, on-the-go customer and inventory insights.
The client's sales team was limited to the data printed ahead of in person sales calls. The client wanted to arm their sales teams with the ability to field any customer questions in real time. The small data team lacked the capacity and deep-level expertise to implement and maintain data modeling best practices within Microsoft Fabric to enable more accurate and effective AI-powered reporting. They sought expert guidance to move beyond the initial setup and accelerate the realization of key business goals – including enabling sales teams with real-time data.
The underlying data in the Microsoft Fabric environment suffered from inconsistent data modeling, technical debt, and a lack of clear standards. Specific technical challenges included:
Inconsistent naming conventions for key fields.
Tables with non-uniform granularity, leading to ambiguity and complexity (e.g., mixing warehouse-level and bin-level records in a single table).
A reliance on basic text editors for development, which made it easier to introduce data quality gaps.
Trility adopted a highly consultative and pragmatic coaching approach designed to meet the client where they were, ensuring the solution is owned and sustained by the internal team.
The team focused on identifying core data modeling gaps (like granularity and inconsistent standards) within three priority data domains.
Work sessions followed this pattern: Trility identified the issues, presented practical solutions, then the client team implemented the fixes, which were subsequently reviewed.
This iterative, hands-on process included coaching, paired programming, and documentation to build long-term internal capability and reduce the risk of falling back into old habits.
Microsoft Fabric: The core data platform where the data modeling and transformation were performed.
Copilot Studio and Copilot: Used to develop and deploy the final conversational AI agent that provides self-service reporting capabilities.
The project delivered tangible, high-value business outcomes, empowering the client's users and establishing a scalable solution.
Conversational AI: The most significant outcome was the launch of an AI-powered, self-service reporting agent. Sales now uses a mobile app with speech-to-text questions (e.g., What products does customer X order most? and Compare year-over-year sales.) and receives instant, accurate answers while on the road.
Business Insights: The solution addressed a long-standing business goal, accelerating the delivery of critical sales insights and expanding the business's data capabilities.
Scalable Best Practices: Trility successfully modeled the priority data domains, removing technical debt and establishing a repeatable, best-practice pattern for integrating, modeling, and connecting data into Copilot. This foundational work provides a solid, scalable base for the client to roll out the solution to additional data domains on their own.
Sustainable Culture: By tailoring the solution to the team’s capacity, the client gained the confidence and knowledge necessary to ensure the longevity of the solution.
