BPO Reduces Bill Of Lading Processing Time By 63%

Industry

Business Process Outsourcing

Product

CMR

Process Automated

Bills of Lading

-63%

Processing Time

up to 80%

Accuracy Capturing Unstructured Data

Client Profile

Founded in 1999, this BPO is a publicly-traded global consulting and outsourcing company with more than 31,000 employees worldwide. It offers insurance, banking, financial services, utilities, healthcare, travel, transportation and logistics services.

Challenges

The BPO’s client wanted to reduce the volume of errors generated via manual processing of its bill of lading documents, manage volume spikes, and accelerate turnaround time. It receives 50,000 documents each day with over 120 fields that needed updating manually. Their existing OCR platform struggled with the sheer variety of the documents arriving from thousands of shippers worldwide. As a result of these issues, there were significant delays in shipments reaching their destinations and, as a result, causing a loss of revenue.

Solution

Cognitive Machine Reading (CMR), with its combination of AI, cognitive technology, Machine Learning (ML) and pattern recognition, ingests and extracts multi-format unstructured data for the client’s billing process. The solution allowed for “drag and drop” business rules, as well as pre-population of more than 1.13 million consignees and over 370,000 addresses. CMR automatically flags missing or unreadable information as exceptions and routes tasks to the right person. Its business rules engine applies rules to enable specific configurations. Each manual intervention improves the accuracy and confidence of CMR’s ML models and multiple reports are available for further analysis.

Results

After CMR was implemented, this BPO organisation saw extraction accuracy levels of up to 80% across the various shipment documents it handled, resulting in efficient exception handling and reduced turn-around time. CMR reduced the processing time by 63% leading to a consistent process flow during peak and off-peak hours.