Polaris transforms cross-border logistics with AI automation

How Polaris built an AI-ready logistics operation

Polaris Transportation Group runs cross border logistics at scale, where speed and accuracy decide everything. To compete in the less-than-truckload (LTL) market, Polaris has embedded automation across customs, order creation, and front line operations—shifting from manual, document-heavy workflows to a more software-driven operating model.

To deliver this at scale, Polaris worked with qBotica through a managed services model built on the UiPath Platform™, with automation led internally by its digital arm, NorthStar Digital Solutions.

The objective was not cost reduction alone, but to remove manual administration, improve accuracy, and build an operating model capable of competing with digital first freight platforms entering the market.

Employees were touching documents two, three, four times and still making a lot of errors, which made the freight movement a poor experience for a lot of our customers.

Dave Brajkovich · CTO at Polaris Transportation Group and NorthStar Digital Solution

Transforming customs clearance

Polaris moves high volumes of cross-border freight where customs clearance directly affects delivery speed and reliability. That process is now largely automated, with unstructured customs documents classified, validated, and processed as they arrive.

Delivered through qBotica’s Document AI service—built on the UiPath Platform™ and powered by UiPath Document Understanding™— customs documentation flows straight through wherever possible, with exceptions routed to the team only when required.

Today, approximately 85% of customs documents are processed automatically, accelerating clearance across time-critical shipments.

As automation matured, teams began owning and refining exception handling themselves, working with qBotica to continuously improve accuracy and straight-through processing.

Order creation embedded into daily operations

Polaris handles large volumes of order requests each day, most arriving through unstructured email. Order creation is now largely automated, with incoming messages read, relevant data extracted, and orders created directly when information is complete.

Requests that require clarification are routed to an exception queue, allowing teams to focus on higher-value decision making. Order creation time has fallen from 15–20 minutes to around two minutes, with approximately 30% of orders now created end-to-end with no employee intervention.

We’ve gone from moving an order from 15 to 20 minutes on average down to two to three minutes, or less in some cases, if the data is really clear and nobody touches it.

Dave Brajkovich · CTO at Polaris Transportation Group and NorthStar Digital Solution

Improving bill of lading alignment at trailer level

In certain customer scenarios, bills of lading arrive with freight rather than through electronic data interchange (EDI). Polaris now processes these documents automatically, reading, collating, and aligning them directly to orders in the transportation management system.

This trailer-level alignment removes hours of manual reconciliation each night and creates a clear audit trail when discrepancies originate from customer provided data. What previously required several people working for five to six hours can now be completed in minutes.

The AI agents can read the document word-for-word and put it into our system. With that, the accuracy matches exactly what the customer has requested. Nobody typed it, hence errors from human processing, mistakes, and disputes are avoided because what we show them is exactly what was sent.

Dave Brajkovich · CTO at Polaris Transportation Group and NorthStar Digital Solution

Connecting drivers to the back office

Automation at Polaris extends beyond the back office. Freight documentation is captured at pickup and delivery through mobile workflows, with documents routed automatically into downstream systems in real time.

This approach triggers customs clearance and invoicing while shipments are still in transit, rather than waiting for end-of-day batch processing. Data quality improves, shipment visibility happens in real time, and delays outside Polaris’ control can be documented clearly. Internally, Polaris refers to this model as the “road to the back office,” using front-line data to drive automation across the entire operation

Building toward AI driven orchestration

Polaris receives thousands of quotation requests each day, largely through unstructured email. While automation already supports initial quote ingestion, manual handling limits response speed and creates the risk of missed revenue opportunities.

Building on its existing automation foundation, Polaris is developing AI-driven orchestration with qBotica’s qubi agent hub, built on the UiPath Platform, to manage quoting at scale. These capabilities are designed to assess incoming requests, capacity, and business policies to prioritize responses, apply dynamic pricing logic, and engage customers more consistently.

One of the things the CEO wanted was to remove this labor-intensive system-to system administration and give our people the ability to do what they went to school for, instead of punching keyboards all day.

Dave Brajkovich · CTO at Polaris Transportation Group and NorthStar Digital Solution

By extending a mature automation backbone with AI-driven orchestration, Polaris is evolving toward a more software-driven logistics operating model—designed to compete in an increasingly digital freight market.