All cases

AI-powered self-service BI and big data platform for the world's leading retailer

Client
One of the largest retailers that operates a chain of hypermarkets, discount department stores, and grocery stores worldwide
Industry
Retail
Services
Product Development, Data Architecture, Data Management, Data Analytics, Data Visualization
Tech
Cosmos DB, MS Azure, Gremlin, JanusGraph, Cassandra, Java, Python, Kafka, React, Redis

Challenge

Data is only as useful as it is accessible. For a company of this size—nearly 3 million internal users spread across multiple store formats—isolated data silos were more than just an IT issue. Disparate systems and data islands were generating inaccurate insights, resulting in poor business decisions and missed market opportunities. The client approached ITRex to fix the foundation.

We took on the following challenges:
Create an easy-to-use custom web portal to enable a 360-degree view of all data sources and data available in any format (Excel spreadsheets, invoices, PDFs, text files, databases, etc.) across the entire organization
Develop a self-service BI platform to provide business users with an opportunity to analyze data and create complex ad-hoc reports
Create a Master Data Repository to serve as a source of truth for all organizational data
Break down data silos and achieve consistency across data sources by collecting and connecting petabytes of data from disparate systems
Provide data cleansing by detecting and correcting incomplete, incorrect, inaccurate, or irrelevant data and then replacing, modifying, or deleting the dirty or coarse data records from original data sources
Manage data security and set up role-based access control
AI-powered Big Data Platform
Big Data Platform

Solution

ITRex designed the platform with three priorities in mind: making complex data queryable without SQL expertise, connecting every major enterprise system used by the client, and keeping sensitive data secure:
Graph data structures to handle complex networks of relationships and resolve deep, multi-hop queries efficiently—something traditional relational databases struggle with at this data volume
A near-natural-language query dialect, significantly simpler than SQL, so non-technical users can search without training
Integration with major data sources and systems: Enterprise Data Lake, Slack, Zoom, Atlassian Stack (Confluence, Jira), Office 365, Active Directory, SAP, Microsoft Exchange, and others
A Report Builder that lets business users query multiple data sources simultaneously and share the resulting reports
Hashtag Autocomplete and Hashtag Search to speed up navigation across a massive data volume and surface the most relevant results
An internal communication function that lets employees send messages to various channels directly from the platform
A simple API enabling faster data collection and supporting custom app development and internal system integrations
Role-based security in graph databases to protect sensitive data by restricting access based on each user's role
Global Retailer solution
Solution for Global Retailer

Impact

The platform is now processing 13 million API requests and hundreds of millions of events per day, with close to 3 PB of data stored in the data lake. Here is what that translates to across the business:
Time-to-insight accelerated by 60–65%. Automating routine reporting and giving 3 million internal users direct, self-service access to previously siloed data cut the data-to-decision cycle significantly. Tasks that previously took days are now completed in minutes—directly improving the organization's agility in responding to market shifts.
A 4–6% increase in revenue driven by faster, data-backed assortment planning, more effective promotions, and identification of hidden demand patterns across hypermarket, discount, and grocery formats—reducing missed market opportunities.
A 20% reduction in overhead costs tied to manual data processing and reconciliation. Eliminating manual reporting workloads, reducing dependency on centralized data teams, and improving data quality all contributed to fewer reconciliation cycles and lower operational drag.
A 10–15% reduction in inventory costs driven by improved demand forecasting accuracy and the elimination of data silos that were distorting stock visibility across the supply chain.
Data analytics teams freed from the majority of routine reporting work—redirecting capacity toward higher-value analysis.
Accurate, relevant data delivered to any internal application built through the API.
We are making in minutes things that our teams have been stuck with for days before.
Product Owner
Global retailer

Latest projects