Maersk uses intelligent automation to enhance invoice processes
Intelligent automation (IA) combines elements of artificial intelligence (AI), robotic process automation (RPA) and business process management (BPM) to help organizations streamline and enhance operations. It aims to transform business processes into autonomous systems which are capable of learning, adapting and making efficient decisions independently.
According to recent academic research, IA is being increasingly adopted in a wide range of customer-facing service settings, with shipping industries adopting IA systems to manage their processes.
Ahead of All Access: Intelligent Automation 2024, PEX Network spoke to speaker Vineet Mehra, senior platform product owner at Maersk and PEX Network Advisory Board member, to explore how the shipping and logistics giant uses IA to enhance operations.
How Maersk is applying IA to enhance operations
“We’re using IA to improve our trucking services supplier invoice processes,” Mehra says. “Around 80 percent of our invoice volumes are PDF copies. We are implementing an IA-ML driven solution that reads the documents, extracts data and transfers it back to our ERP system to post the invoice automatically.”
Some of the proof points Maersk tested were how well the technology utilizes the patterns that vendors are following and how well it can extract the information from the PDF so that human capital can be used in doing something better, he adds. “Traditionally, we had an army of people who looked at these PDF invoices – not just once or twice but, in some cases, up to 10 touchpoints to arrive at final processing. They had to go back and forth, read and extract the data and put it into the system and also work with operations teams to resolve issues.”
If something failed, they’d then need to go back to the vendor again. “So we are trying to reduce the touch points, and the main measure is how many times we are able to process an invoice first time right. The second thing we measure is the end-to-end automation rate, which measures the number of invoices that can be processed without any human touches at all.”
The ultimate goal is to pay truckers faster to claim early pay discounts, or to pay as per agreed payment terms, Mehra says. “We are not always able to pay on time with purely manual-focused processes. It’s at the initial stages and we are only at around 60 percent coverage globally, but the plan is to be running at 90 percent by the end of this year/into next year by gradually increasing adoption.”
The challenges of IA
The biggest challenges are integrating the technology with Maersk’s multi-layer ERP system and driving change across the organization, Vineet says. “We also have to standardize our internal processes, and the way we run processes differs from country to country, region to region. We did a proof-of-concept last year where we tested and went back to the business stakeholders, showing them the value IA brings in streamlining end to end flows. Every time we were rolling out in a new country or market, it was important to collect feedback because the process dynamics are different country by country.”
The main thing is to really understand the processes, standardize as much as possible and set early expectations. There will be a lot of things that need to happen regarding Tax compliance and local regulatory adherence, Mehra adds. “We decided we need to have a dedicated set of users to focus on machine training invoice patterns, because we want to make sure the system learns from a particular pattern over a period of time.”
All in all, it is a journey to provide better user experience by integrating IA solutions and streamlined business processes to ensure cost controls and satisfied suppliers, Mehra concludes.