An AI-ready data platform on Microsoft Fabric

Check out how we centralised data and built a modern data platform ready to implement AI at scale for the leader in the US online insurance market, Nsure.com.

Nsure.com is an innovative platform whose story is as inspiring as the beginnings of Netflix. The founders, looking for the right insurance for themselves, saw the need for an intuitive online platform that would let customers buy it without leaving home: from preparing a calculation, to comparing quotes, to finalising the transaction. And so they built it. Since then, the company has grown rapidly, becoming not only a leader in the US online insurance market, but also a pioneer in basing its search engines on the latest technologies, including artificial intelligence and machine learning. It now operates in all 50 US states, taking the classic insurance office into the digital world.

Azure data warehouse, Dataverse, Cosmos DB, Twilio

// Solutions

Microsoft Fabric Implementation

Data Warehouse, Lakehouse, Data Fabric & Data Mesh

Data Platform Migration & Modernisation

Data Integration & Consolidation

Data Agents & Operations Agents

Semantic Models & Data Ontology (Fabric IQ)

Data Visualisation & Executive Reporting with Power BI

Reports Migration from Excel, Tableau & Qlik to Power BI

// Industry

Insurance

// Stack

Microsoft Fabric

OneLake

Power BI

Azure

The challenge

Slow, unreliable reporting, data scattered across systems and operations done by hand. Nsure.com needed one central platform ready for more AI components.

As it grew, Nsure.com recognised the need to manage and centralise large volumes of different types of data. As experts in Data & AI and Microsoft technologies, we were chosen to implement a modern data platform ready for the AI era.

What prompted the need for a modern, central data platform?

Nsure.com’s goal was to prepare the platform for its expansion with more AI components, to deliver even more personalised offers to its customers. To do so, we faced the following challenges:

Ineffective reporting system. Solutions based on the Azure platform did not meet expectations in terms of speed and reliability.
Data dispersion. The information was in different systems, making it impossible for teams to work on the same data.
Manual processing of information. Some operations were carried out manually, which consumed the teams’ time.

Our approach

First the current state and the needs of every department, then the target architecture and a phased schedule. Implementation ran alongside the existing reporting, with workshops throughout.

Analysis of the current state
We started with a detailed review of the existing data infrastructure to understand what changes were needed. We analysed the client’s data warehouse architecture on the Azure platform and the ETL/ELT processes involved in processing the source data. We then met representatives of all departments that would benefit from the data platform (reporting, data science and AI) to learn and fully understand their needs.
Defining the target state
Based on the analysis of the current state, we developed a concept for the future platform. Setting end goals: we defined the project goals and the final architecture of the new platform. Task breakdown and milestones: we prepared an implementation schedule, dividing the implementation into phases.
Start of implementation
During the project, we were keen to interfere as little as possible with the client’s current infrastructure and analytical solutions, to ensure continuity in reporting and data analysis. To ensure a smooth transition to the new platform, we met Nsure.com representatives regularly to adapt the platform to their needs in real time and transfer relevant expertise on an ongoing basis.
Training and workshops
We wanted Nsure.com to take advantage of the full potential of the solution we implemented, so throughout the project we regularly shared knowledge with the client. We ran workshops that allowed Nsure.com employees to quickly gain the skills needed for the new platform.

Solution concept

Sales, marketing, Twilio and quoting data flow automatically into one Microsoft Fabric lakehouse. Power BI models were tuned, and insurers got reports on their own performance.

Synchronising data from key business applications

We synchronised sales data from Power Apps applications stored in Microsoft’s Dataverse database, marketing data stored in Azure Cosmos DB, and data from internal process management applications in the organisation, such as Microsoft Shifts.

Transferring data to Microsoft Fabric

We automated the process of extracting data from the applications and then transferred it to a Modern Data Lakehouse based on Microsoft Fabric.

Data integration and automation with Twilio

Twilio is a Communications Platform-as-a-Service (CPaaS) that Nsure.com used to communicate with customers. We integrated it with the data platform, which automated data retrieval and significantly reduced the time the data science team needed to prepare reports.

Quote automation data integration

We integrated the information generated when preparing personalised insurance quotes into the data platform. As a result, employees can monitor the performance of the quoting systems and optimise them.

Optimising the performance of Power BI semantic models

We advised the data science team on optimising the performance of semantic models, which improved the overall experience of using Power BI.

Access to Power BI reports for insurers

Insurers working with Nsure.com gained access to Power BI reports on their own performance for internal analysis.

Results

By implementing a platform based on Microsoft Fabric, Nsure.com has achieved a number of benefits:

A single, up-to-date source of data. Collaborating teams gained access to a single, consistent database, which made collaboration and strategic business decisions easier.
Reduced operating costs. Using Microsoft OneLake allowed us to reduce how often data is exported and duplicated, and to take advantage of shortcuts in Microsoft Fabric, which reduced costs.
Process automation. Automating previously manual processes and integrating the internal systems in use has freed up the teams’ time.
Readiness for AI development. We have transformed the way teams access, manage and act on data with a single modern data platform ready to plug in AI-based solutions.

Data analysis is the foundation of our business – we make key business decisions based on it. The implementation of Microsoft Fabric has significantly revolutionised our approach to analytics, enabling us to create a unified data lake and integrate all sources, which has significantly accelerated data processing. As a result, analysts working in both Power BI and notebooks can effectively use common models, resulting in more precise analysis and visualisation of data in Power BI.

Wojtek Gudaszewski

Co-founder and COO, Nsure.com