Overcoming Operational Data Bottlenecks in Rental Billing

For industrial gas and welding distributors managing assets like cylinders and bulk tanks, rental billing represents a vital share of revenue. However, extracting actionable insights from operational systems historically required submitting requests for custom SQL Server Reporting Services reports or raw data exports through the support team at Datacor.

Customers often waited days or weeks for static outputs that answered only the initial query. This manual workflow prevented business teams from quickly iterating on pricing or fleet utilization strategies, leaving potential revenue uncaptured across millions of tracked assets.

Embedding Generative BI and Dashboards into TrackAbout

To remove these technical hurdles, <a href="https://www.datacor.com/products/trackabout">TrackAbout</a> integrated embedded analytics capabilities powered by <a href="https://aws.amazon.com/quicksight/">Amazon Quick Sight</a>. The enhanced platform equips business users with pre-built, context-aware dashboards displaying fleet utilization rates, billing exception summaries, revenue recovery trends, and customer-level asset aging.

Additionally, the solution features a generative business intelligence search bar powered by Amazon Q in QuickSight. Users can type plain-language prompts to analyze rental performance metrics, review unreturned cylinder totals, or calculate average rental rates across specific customer accounts without filing IT tickets.

Designing the Automated Cross-Cloud Data Pipeline

Behind the self-service interface lies a structured architecture spanning data ingestion, transformation, and storage. TrackAbout operational data resides in Azure SQL databases, from which an Azure Data Factory pipeline extracts source tables as Apache Parquet files and streams them into <a href="https://aws.amazon.com/s3/">Amazon Simple Storage Service (Amazon S3)</a>.

The pipeline handles both full and incremental loads using change tracking to move only modified rows. The arrival of new files in storage triggers an orchestration layer built on <a href="https://aws.amazon.com/lambda/">AWS Lambda</a>, <a href="https://aws.amazon.com/glue/">AWS Glue</a>, and <a href="https://aws.amazon.com/dynamodb/">Amazon DynamoDB</a> to manage processing jobs and single-threaded execution.

Ensuring Data Integrity and Multi-Tenant Security

AWS Glue jobs merge incremental updates into Apache Iceberg tables located on Amazon S3. An AWS Step Functions workflow validates row counts and constructs final reporting tables before refreshing datasets within <a href="https://docs.aws.amazon.com/quicksuite/latest/userguide/spice.html">Amazon Quick Sight SPICE</a> to ensure users view current operational metrics.

Sundar Kuppuswamy, Chief AI Officer at Datacor, emphasized that the underlying foundational engineering was critical to the project's success. He noted that cross-cloud ingestion, row-count validation, and a verifiable tenant-isolation model were essential before introducing natural language querying to revenue data managed on <a href="https://aws.amazon.com/">Amazon Web Services</a>.

Sources

• AWS Machine Learning BlogHow Datacor built self-service rental analytics with Amazon Quick Sight