Document Processing · IT Operations Automation
Pilot Company needed a better way to support fuel bill-of-lading document processing and billing automation because hundreds of handwritten form formats defeated OCR and RPA approaches, leaving a critical fuel billing workflow slow, error-prone, and expensive.
Reported impact
Pilot Company improved fuel bill-of-lading document processing and billing automation with measurable results including 99% fueling efficiency from pump to profit, 90% decrease in overall processing costs, and Intelligent document processing at 98.6% accuracy with Databricks Agent Bricks, while giving teams a faster path from governed data to production decisions.
Pilot Company uses Databricks Data + AI Platform to support fuel bill-of-lading document processing and billing automation. The implementation replaced a legacy state where hundreds of handwritten form formats defeated OCR and RPA approaches, leaving a critical fuel billing workflow slow, error-prone, and expensive. Handwritten bills of lading and fuel delivery photos move through production data pipelines and land in Delta Lake, giving teams one trusted data foundation. Unity Catalog governs access, lineage, and compliance across sensitive datasets. MLflow, Agent Bricks, and Amazon Bedrock operationalize models, agents, and analytics that deliver structured billing records and automated billing decisions. Pilot Company improved fuel bill-of-lading document processing and billing automation with measurable results including 99% fueling efficiency from pump to profit, 90% decrease in overall processing costs, and Intelligent document processing at 98.
Handwritten bills of lading and fuel delivery photos land on a shared platform, pipeline services clean and standardize them, governance controls enforce trusted access, and AI, model-serving, and analytics services deliver structured billing records and automated billing decisions.
One of our biggest roadblocks was evaluating models,
We moved to a multi-modal model and had controlled access through Databricks. Feeding the model the actual photos of the BOL rather than extracted text boosted our accuracy all the way up to 94.3%,
We think it’s because of the model’s ability to preserve spatial reasoning — the layout of the checkboxes and form sections. These things previously were lost.
We’ve tried traditional robotic process automation (RPA), but the forms varied so much it wasn’t feasible,
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
Databricks Sample Sep 24, 2024 Hapag-Lloyd Hapag-Lloyd needed to reduce manual work in corporate audits because producing findings and executive summaries from bullet points was time-consuming and inconsistent. 66% Agent Bricks · Databricks Data + AI Platform · Databricks Model Serving 
Google Cloud Sample Modified · Oct 9, 2026 Porter Porter (Smartshift Logistics Private Limited), a leading logistics platform based in India, orchestrates on-demand mini-trucks, tempos, and two-wheelers across 50+ major cities. When order volumes doubled, Porter’s… Achieved 30% zero-touch driver verification automation on Gemini platform BigQuery · Cloud Composer · Cloud Pub/Sub Databricks Sample Jan 21, 2026 Mirakl Mirakl needed a better way to support generative AI supplier catalog onboarding for ecommerce marketplaces because strict governance rules and siloed data teams slowed marketplace analytics and made supplier catalog… 30x faster iteration and delivery across data and ai teams Agent Bricks · Databricks Data + AI Platform · Lakeflow Jobs 