This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
July 2026
By Jeff Stagl, Managing Editor
Technology advances in the freight-rail industry have led to record-breaking safety gains and helped railroads boost safety, efficiency and reliability.
That’s what Association of American Railroads President and CEO Ian Jefferies emphasized to the Senate Committee on Commerce, Science and Transportation’s Subcommittee on Surface Transportation, Freight, Pipelines and Safety June 9 during a hearing that examined how technological advances are driving transportation innovation.
The modern rail network increasingly relies on artificial intelligence (AI)-assisted inspections, machine vision, predictive analytics, advanced sensors and real-time data to help identify risks earlier, improve safety and enhance service reliability, he stressed in his testimony.
“And the results are clear. According to the Federal Railroad Administration, 2025 was the safest year on record across key measures, including derailment rate, equipment- and track-related accidents, and ... employee injuries,” Jefferies said. “This progress reflects sustained investment, employee expertise and continuous innovation.”
The future of rail safety depends on detecting issues before they become failures, and Class Is are employing a number of advanced detection technologies, he added.
For example, CSX uses a system with GPS, cameras, radar and lidar to monitor work zones to automatically stop equipment movements when risks are identified; Norfolk Southern Railway deploys AI-powered imaging portals that scan rail cars at track speed and flag defects in near real time; and Union Pacific Railroad uses machine vision and AI to analyze large datasets, detect patterns and predict maintenance needs months in advance.
“Other railroads are advancing similar innovations. CN uses automated inspection systems and ground-penetrating radar, while CPKC applies optical technology to detect early-stage defects,” said Jefferies. “Across the industry, technology is expanding both the reach and precision of inspections.”
What CPKC strives to keep expanding is the adoption of transformative technologies in many aspects of its organization.
The ultimate goal is to accomplish a host of objectives by developing, adopting, enhancing or leveraging many innovations. Among them: bolster operations, enhance service performance/reliability, broaden efficiencies, increase velocity, improve safety, control costs, strengthen workforce development, upgrade maintenance processes and foster better asset management.
From autonomous rail-car inspections to predictive analytics to AI, innovation is crucial to the railroad’s future, CPKC executives say.
“Technology is a way forward. It’s a competitive advantage, and not just in rail,” says Assistant Vice President of Operations Technology Kyle Mulligan. “There is diversity in what we are doing, focusing on implementing and recognizing technologies.”
The railroad also aims to become much more predictive than reactive when it comes to infrastructure and rolling stock maintenance, and much more tuned into how innovations can help employees better perform their jobs, says Vice President of Engineering Tom Bourgonje.
“We have great railroaders, real boots-on-the-ground railroaders. We want to use technology to develop things that are meaningful to these great people,” he says.
One major challenge with meeting the lofty technological goals: managing the huge amount of data collected by various systems, sensors and devices. The sheer volume of images and data generated is vast, and gaining insights into data is crucial, says Mulligan.
For example, two test cars that analyze water sources collect terabytes of data every day, says Bourgonje.
“That’s terabytes, as in plural,” he says. “We are finding ways to manage that data that’s meaningful, so we don’t waste all that data.”
CPKC Director of Track Systems John Furlong stresses that cloud computing is needed to process the tremendous amount of data generated by trackside technologies.
“If you tried to process it on a laptop, it would catch fire,” he says.
For the past eight years or so, technological advances and data collection have instilled more intelligence into the railroad, says Shailesh Yerram, CPKC’s AVP of business intelligence and chief information security officer.
But it’s vital to “connect the dots” so that intelligence makes its way from a sensor to a user and all the way to the back end where trains are designed, he says. And the incorporation of AI is critical.
“We are finding ways to turn AI data into models and models into intelligence,” says Yerram. “If you leverage data to better understand operations, that is the secret sauce.”
That, in part, means “getting AI right” to find the right balance, he adds. And CPKC is striving to do just that.
In May, the Class I created a center of excellence for data analytics and AI. The center is staffed by three data scientists overseen by a manager.
“We are trying to do more with technology on our own, to create expertise and our own experts,” says Mulligan.
The data scientists are undergoing a lot of training — such as becoming certified as conductors — to learn the railroad business and better understand AI’s role in it, he says.
“They will look at where it makes sense to use AI, the costs versus benefits. It’s sort of an acid test,” says Mulligan.
The center’s main goal for now is to try to reduce train accidents by 15% within a 12-month period. If that objective is met, the goal will be taken further, says Mulligan.
The railroad currently uses AI to identify track buckles, determine rail stress, and interpret generated images and data. CPKC aims to be at the forefront of predictive rail analytics through wayside detectors, rolling-stock sensors and various systems.
The railroad strives to leverage analytics to focus infrastructure maintenance projects where they have the most impact, employing a proactive approach to help improve reliability, increase train speed, and reduce asset and terminal dwell times.
And it takes a lot of trackside systems and devices to generate useful data and gain advantages from analytics. CPKC’s current wayside technologies include:
CPKC also has developed lower-cost hot-box detectors that monitor the temperature of passing train wheel bearings. As the cost of the detectors drops, the railroad can install more of them in the field to collect more data, says Mulligan, adding that the Class I now has 130 of them along track.
Another recent development involving rolling stock: the installation of StarLink satellite network devices on locomotives for monitoring in dark territory. The railroad has continued to install the devices over the past year.
StarLink provides real-time telemetry information, enabling CPKC to make better decisions based on asset health and monitor locomotives for operator train-handling exceptions, says Mulligan. The technology offers a pathway toward the next generation of train control instead of installing cell towers across the country for connectivity, he says.
The railroad’s innovation team and internal information technology department are working to design the integration of StarLink into existing locomotives, he says.
In addition to StarLink-equipped locomotives, there are other high-tech pieces of rolling stock equipment traveling along CPKC’s network, including a track evaluation train (TET).
The three-car train — which includes an instrumented track geometry car, power car and accommodation car for the track evaluation crew — inspects all track along CPKC’s 20,000-mile system in Canada, the United States and Mexico. On an annual basis, the TET collects data along more than 300,000 miles of track.
An engineering department team evaluates all the information generated by the TET to plan and execute repair and maintenance work on track.
CPKC also employs six autonomous track geometry measurement system box cars that serve as track evaluation rail cars. The cars operate in revenue service, including on intermodal trains, to collect data, which is wirelessly communicated to local track maintenance teams within 48 hours of an inspection/evaluation. Any urgent track defects are validated prior to advising field workers to address an issue.
Moreover, CPKC uses rail-flaw detection hi-rail trucks equipped with state-of-the-art ultrasonic and induction test systems to detect internal rail flaws. AI is used to analyze the data gathered by the test equipment to identify potential flaws based on small changes from previous tests. The trucks also can help identify missing nuts or bolts and cracked bars.
There are other high-tech aspects to CPKC’s track inspection arsenal, as well. They include:
For the past year, CPKC also has used its various systems and collected data to create a track stress index. The formulated indices help prevent track buckles, says Bourgonje.
“We find the problems and then have people in the field put eyes on those spots,” he says. “We are trying to process data in a way that a foreman can understand, so they can better react in the field.”
CPKC currently is measuring rail temperature at 60 mph to identify force in rail.
“We want to be able to determine the neutral temperature of rail, which is tough to do and a long shot without cutting rail and pulling spikes. We think we can get there,” says Bourgonje.
CPKC also is performing 3D assessments to gain useful images of various track components. Image analysis can help identify spikes that are too high, spikes that are off a tie or missing anchors, says Furlong.
A 3D tie assessment system can grade ties on equipment moving 60 mph. In addition, a rail-web system generates images to provide an inventory of rail, such as its weight, manufacturer and the year it was rolled.
“We also get an inventory of welds and joints. It’s a good inventory overall —10 years ago, that was just a pipe dream,” says Furlong.
As CPKC continues to employ more systems, sensors and devices, one challenge is ensuring a system’s optical devices on cars are always functioning properly. The railroad is working with vendors to find ways to keep the optical devices clean from mud, dirt or snow, says Furlong.
“If there’s an issue, we have to take the car out of service to clean the optical device,” he says. “We could use an air knife that provides constant air or add a self-cleaning system.”
While juggling the impacts of more technologies employed over time, the railroad is considering other innovations, too. For example, CPKC is analyzing an autonomous ground-penetrating radar system that could be installed on a box car to monitor ballast condition, such as the location of any voids or water pockets.
“We plan to try it on one box car. We are looking for technologies to help us be more flexible and nimble when moving equipment around the system,” says Furlong.
In addition, CPKC aims to keep employing technologies in more ways. For example, a geohazard management system helps the railroad monitor waterbodies as well as identify potential avalanches, landslides bank erosions and wildfires.
The system is especially helpful because the climate is so erratic now, says Bourgonje.
A water body hazard monitoring system (WHMS) uses AI, satellite data, synthetic aperture radar and multispectral imagery to detect water-related hazards along CPKC’s network and issue alerts on potential washouts. The system was developed several years ago by consulting/engineering services firm Tetra Tech and the railroad’s geotechnical engineering team.
WHMS scans more than 1.2 million bodies of water along more than 12,600 miles of track (in Canadian Pacific’s former network) and generates satellite images, providing near real-time visibility across the network.
By early July, CPKC expects to deploy the second version of WHMS, which will enable the railroad to analyze water bodies along Kansas City Southern’s former network in the southern U.S. and Mexico, says Bourgonje.
For the past six months or so, the Class I also has been employing two test cars with lidar to create a data risk index of water sources. It will take about a year to complete indices for more than 10 million water bodies, both above and below ground, says Bourgonje.
“The index will consider all factors and determine how water will hit a track,” he says.
While dealing with flooded track can be aggravating for CPKC, confronting manual customs processes can be frustrating for shippers.
That’s why the Class I last year developed a web-based Customs Logistics Easy and Rapid (CLEAR) system, which not only is designed to streamline customs processes, but also boost efficiency, enhance the customer experience and attract new business.
More than 50 team members from diverse CPKC departments contributed to the system’s development. CLEAR currently is being used by customers at the Port of Lazaro Cardenas in Mexico and Port of Saint John in Canada.
Rather than manually input data from extensive shipping documents to meet customs and other cross-border requirements — which is a time-consuming process — customers can use a web-enabled CLEAR portal to more easily and quickly enter information.
The system leverages robotic process automation and intelligent document processing to extract and organize required information automatically.
“Instead of making phone calls back and forth, CLEAR uses AI and speeds up the process,” says Caryna Pinheiro, CPKC’s AVP of applications and digital services and assistant chief information officer. “Customers have said they like the efficiency.”
CLEAR has helped reduce customs processing times from weeks to hours while providing customers real-time visibility of their shipments.
CPKC is considering an expansion of CLEAR to other ports and corridors and is exploring ways to leverage the system as a foundation for improving the customer experience across its network.
“You can sometimes be disabled by technology, which can be a problem. You need diversity of thought,” says Pinheiro.
Meanwhile, CPKC is working on a service design modernization that involves technological improvements. The aim: to embed intelligence into the platform, says Yerram.
“We want to gain more insight and not just throw data to the user,” he says. “I think we will have the first phase of this in the fourth quarter, and then incrementally add factors to it.”
A hurdle that needs to be cleared for the project is tapping the vast institutional knowledge in the field.
“We call it ‘tribal knowledge,’ and we need to codify that knowledge from people who know how to do things in their head,” says Yerram. “We need to make that knowledge available to the end user.”
Perhaps that undertaking can best be categorized as easier said than done. As is CPKC’s goal of becoming the best Class I when it comes to technological advances.
For Bourgonje, the railroad has made notable strides in that regard. He had previously served CN for more than 40 years, lastly as AVP of engineering when he retired in 2022. But later that year, Bourgonje came to Canadian Pacific, and he’s witnessed a lot of positive technological changes since.
“When I joined them four years ago, I think CP was lagging in technology. But we have had a lot of buy in and ‘get em to try it,’” he says. “And now, I think we have gone from lagging to leading the pack.”
Related Topics: