Powered by AI, automated MRFs can revolutionise the business of recovery
Efficiency and improvement advances in computer vision-based sampling will transform future recycling operations.
According to the UN’s latest research, global municipal solid waste production is projected to reach 3.8 billion tonnes annually by 2050, a staggering 1.5 billion tonnes more than in 2023. At the same time, demand for high-quality recyclates is growing with stronger packaging regulation, which comes with its own sampling demands.
To cope with this tsunami of waste, the world’s largest plant builders have embraced digitisation and artificial intelligence (AI) to shape the next generation of fully-automated facilities.
Facility wide automation
While early adoption of smart robotic arms showcased the potential to automate sorting challenges, today’s industry leaders are moving towards end-to-end automation with AI powering the data capture necessary to automate facility wide.
This evolution isn’t merely about replacing manual labour with robots; it’s about leveraging AI-enabled data and analytics to optimise every facet of waste recovery — from automating waste infeed analysis to streamlining compliance reporting.
After all, even the best automated sorting solutions won’t increase the purity and value of recyclables if contaminated feedstock comes into the facility. Therefore, waste data that can be gleaned from the beginning of the waste stream can be extremely valuable as well. AI-powered cameras or sensors within this part of the process can identify waste composition to rebalance the materials a MRF is getting from suppliers.
This facility wide AI integration allows plant builders to interconnect these tools, establishing fully automated, end-to-end sorting systems that are continuously monitored and adjusted using the data AI provides.
Central to this shift is waste intelligence — the actionable insight gathered from detailed waste data that drives decision-making and operational efficiency. Unlike the approach seen during the robotics era, when hardware investments often outpaced software-driven insights, modern AI applications prioritise comprehensive data analytics.
This approach not only enhances sorting accuracy but also ensures facilities meet stringent regulatory demands for material purity, such as the EU’s 95%+ purity requirements for food-grade recyclables.
Why automation will boost waste management’s bottom line
Growing demand for high-quality recyclates like rPET necessitates MRFs capable of maintaining throughput without sacrificing sorting accuracy. Manual sorting processes struggle to maintain this balance, meaning automated MRFs that can adapt to changing waste compositions in real time are increasingly sought after.
These facilities ensure precise material tracking and enhance quality certifications—something that’s becoming more important with regions like the EU introducing stricter guidelines.
Early last year, we heard that the sampling demands of extended producer responsibility (EPR) could mean that a quarter of UK facilities need upgrades worth £50M each.
The Environment Agency has indicated that it’s willing to accept reports generated by accurate waste intelligence platforms, which is exciting news for MRFs and PRFs.
At the scale of today’s waste flows, automated material monitoring is about 250 times cheaper than manual sampling, and automated compliance reporting has already been successfully tested in one of the UK’s largest MRFs.
Reducing human contact with waste to ease the labour shortage
The growing labour shortage is still a common theme in my conversations with waste and recycling sector leaders. Even if it was cost-effective to scale up manual processes, the industry would struggle to fill those new roles.
With automation, we have an opportunity to reduce the reliance on manual sorting and sampling, which also presents a health and safety risk to people on the “front line” of waste management.
Tomorrow’s waste professionals will be managing operations from a centralised data dashboard, rather than a conveyor belt.
Putting recovery facilities at the heart of circularity
Beyond resource recovery, the data generated at fully automated facilities is driving a huge increase in our understanding of the world’s post-consumption material.
The “dark abyss” of post-consumption resources has meant that recovery facilities receive waste flows contaminated by non-recyclable packaging design, delayed regulation, and inconsistent collections.
As the saying goes, “a rising tide lifts all boats” – by sharing the detailed waste data gathered at automated facilities with stakeholders across the value chain, we can ensure the material they receive has already been designed and regulated with circularity in mind.
With improvement in the quality of input material, we’ll see higher recycling rates, and a healthier flow of recyclates to fuel the circular economy .
Gaspard Duthilleul is Chief Operating Officer at Greyparrot
