AI-driven solution reduces shipment categorization errors in iOS app

As businesses look to AI for practical solutions, Shift Lab recently delivered an impactful AI-driven feature for a regional shipping company. Faced with ongoing shipment categorization errors, we integrated a machine learning model into their iOS app, dramatically reducing errors and streamlining their operations. Here’s how our AI solution helped solve a persistent logistics challenge and improved overall efficiency.

The Challenge: Persistent Driver Input Errors

A regional shipping company faced a recurring challenge with shipment categorization. Drivers, tasked with providing real-time status updates on their pickups and deliveries, occasionally miscategorized shipments—specifically when toggling between “freight” and “cartage” options. Despite extensive training, this issue persisted, causing confusion in the company’s customer portal, which displayed inaccurate information. The manual toggling system was prone to human error, and while the percentage of incorrect categorizations was small, its impact on customer experience and operational accuracy was significant.

The Solution: AI-Powered Image Classification

To address this issue, Shift Lab introduced an AI-powered solution within the company’s existing iOS app. The AI model, built using Apple’s CreateML, was designed to classify shipments automatically by processing photos of the bill of lading—a document that holds critical information about each shipment. Instead of relying on drivers to toggle shipment categories, the AI reads the bill of lading image, identifies whether the shipment is “freight” or “cartage,” and assigns the correct category with 86% accuracy.

Model Training and Implementation

The AI model was trained using a real dataset consisting of bill of lading images captured by drivers in the field. We structured the training data into categories—freight, cartage, and “other”—using over 6,000 document images to ensure comprehensive model training. By including an “other” category, the AI was able to differentiate between valid and irrelevant documents, increasing its reliability. The model was configured to only make automatic categorizations when it achieved a 95% confidence threshold, ensuring that the system prioritizes accuracy and reduces the likelihood of incorrect classifications.

The Results: 90% Reduction in Errors

Since the AI implementation, the shipping company has seen a dramatic reduction in shipment categorization errors—down by 90%. The solution has not only improved operational efficiency but has also significantly enhanced the accuracy of real-time updates displayed in the customer portal, leading to better customer satisfaction. With the AI seamlessly integrated into the iOS app, the company’s logistics process has become more reliable and less dependent on human error.

Conclusion: AI Driving Operational Excellence

This AI-driven solution exemplifies how machine learning can optimize logistics operations and solve real-world problems. By leveraging AI to handle complex document classification, Shift Lab helped the company automate a critical part of its workflow, reducing errors and improving the customer experience. This success story highlights the transformative power of AI in streamlining logistics, improving accuracy, and driving operational excellence.