Sign up for the Design News Daily newsletter.
The Mail Must Go Through, and AI is Here to Help
The U.S. Postal Service is using artificial intelligence to modernize the procedures and services of this sometimes disrespected institution.
December 31, 2022
3 Min Read
The U.S. Postal Service is investing heavily in AI and machine learning to update its infrastructure and improve the delivery experience for millions of Americans.Image courtesy of Allen Creative / Steve Allen / Alamy
Is this age of e-mail, instant messaging, and social media, traditional snail mail is viewed as one of the remaining bastions of antiquated practices, with the likelihood that unclear or misaddressed mail will arrive at its destination late, or now at all. Still, the U.S. Postal Services processes on the average of well over 100 million pieces of first-class mail daily, and the institution is by no means standing still in trying to update its operations.
During a session at the recent AI Summit New York City titled, “Delivering for America: Transforming one of the world’s largest technology infrastructures,” Pritha Mehra, Chief Information Officer & EVP - United States Postal Service, gave a glimpse of the U.S. Postal Service’s efforts, which leverage the use of AI and machine learning.
“We are seeking to transform the postal service network from one in financial crisis to a best in-class mail and shipping operation,” Mehra told a packed audience. To Mehra, this meant improving and driving the customer service experience, improving operating efficiency, and improving security.
A Herculean Task
Updating the Postal Service’s complex infrastructure is no easy task, given the many disparate legacy systems that are in place and the sheer size of the operation.
But progress is being made. The U.S. Postal Service is considered the global leader in optical character recognition (OCR) technology, with machines reading nearly 98% of all hand-addressed letters and 99.5% of machine printed mail.
The age-old problem of mail getting lost and misrouted has by no means been completely solved, but Mehra said the postal service is working hard to mitigate these issues.
In recent years, the U.S. Postal Service has invested millions of dollars on a machine learning platform that provides analytics, uses diagnostics to help reduce failures, and forecast future mail delivery performance. Mehra said the predictive analysis platform should give customers a more accurate projection of how long a letter or package will arrive at its destination.
Mehra added that the Postal Service has integrated the machine learning platform into a virtual call center to interact with customers and answer questions.
Integrating AI into Delivery
The image of the postal service worker navigating through all types of difficult weather conditions to deliver mail and packages still holds true, as robots are so far not taking over this time-honored task. But the postal service worker now has more help as AI is also playing a much bigger role in the delivery ecosystem, according to Mehra.
“We are harness trip information with the IoT to enable real-time visibility into the logistics network,” Mehra said. “Deliver vehicles have IoT sensors.” Just as Amazon uses telematics to keep tabs on when packages are delivered, the U.S. Postal Service can do likewise do the same delivering packages and high-priority letters.
With concerns for theft omnipresent, Mehra added that the Postal Service is also AI and machine learning to identify customers and reduce instances of fraud.
Upskilling its Staff
The numerous AI algorithms used to upgrade the Postal Service’s network have been developed through internal staff and outside contractors. But Mehra expects the Postal Service to offer greater opportunities for technology professionals as AI and machine learning continue to be adopted. She cited the need for data architects, data scientists, cloud specialists, and of course AI and machine learning experts.
“We will invest in infrastructure optimized for AI and prioritize upskilling and training staff. We will automate processes to rapidly scale AI, and focus on ethics, transparency, and privacy.”
COVID test kit delivery
If one has any doubt whether the Postal Service has the muscle to handle a critical operation on a massive scale, Mehra said one only must look at the COVID-19 test kit delivery during the pandemic. “The technology team had to set up a website to process orders and prepared for delivering at least 20 million test kits. We set up the ordering site on the cloud, relying on cloud-native and edge capabilities to handle the processing. “
Mehra added, “This is an example of how the U.S. Postal Service is equipped to do an operation of any size and scale.”
The U.S. Postal Service recently embarked on another campaign to deliver COVID-19 test kits nationwide.
About the Author(s)
You May Also Like
Building a Wall/Desk Clock With a Customized Arduino and NeoPixel DisplaysFeb 26, 2024|8 Min Read
How to Build a Better Control System Utilizing Microservices and APIsFeb 26, 2024|6 Min Read
Intel Seeks to Grab More of Semi Foundry BusinessFeb 26, 2024|4 Min Read
Revived Geneva Motor Show Comes to Life for 2024Feb 26, 2024|8 Slides