Those of us in the RecTech industry have reached a consensus; the circular economy must become a reality and the path forward is led by innovation in technology.
In the race to revolutionise plastic packaging detection and sortation, two technologies stand out: Polytag, with its UV tag detection system developed in partnership with Pellenc ST, and Greyparrot, which leverages artificial intelligence through its newly-launched DeepNest database. While both aim to improve recycling outcomes, their approaches – and implications – offer different pathways forward.
Polytag: Precision Through Standards and Light
Polytag is championed by GS1, the international not-for-profit organisation behind standardised barcodes and QR codes. This is because Polytag’s system uses UV tags printed on packaging with invisible fluorescent ink. These tags encode a GS1-compliant data matrix, embedding a rich array of packaging attributes – such as consumer packaging variant (CPV), melt-flow temperature, food-grade status, plastic generation, and GTIN-level brand detail.
When packaging passes under a Polytag detection unit, the UV tag fluoresces brightly, and the data matrix is read instantly. This enables absolute certainty about the packaging’s composition and provenance.
The system is designed for open participation: any manufacturer can generate a compliant data matrix, any recycler can install a detection unit, and any ink supplier can develop UV inks that meet open fluorescence standards. This ecosystem model – built on GS1 open standards – encourages widespread adoption and innovation while avoiding potential technology dependencies.
Greyparrot: Camera-led identification
Greyparrot is a company focused on applying AI to waste management, specifically using computer vision to analyse waste streams on conveyor belts in recycling facilities. Their core technology, the Greyparrot Analyser, aims to identify and categorize various waste objects in real-time, providing data on material type, estimated value, and other attributes. This approach is intended to offer greater visibility into waste composition compared to traditional manual sampling methods.
The company has also introduced Deepnest, an AI platform designed to supply brand owners with data regarding the recyclability performance of their packaging within actual waste management systems. Currently, we do not have data on the accuracy of this tool.
While AI in waste management holds promise for enhancing sorting accuracy and data collection, its implementation involves considerations such as the initial investment in technology, the quality and representativeness of data used to train AI models, and the challenges of integrating new systems with existing infrastructure. The effectiveness of these AI systems can be influenced by varying waste compositions and the need for adaptable solutions, yet they are generally seen as contributing to efforts to improve material recovery and inform more sustainable product design.
Key Considerations in the Technology Comparison
1. Certainty of Composition: Data Matrix vs AI Analysis
Greyparrot’s AI system analyses images of packaging to determine material type. However, AI systems face inherent limitations when accessing critical attributes like CPV or melt-flow temperature – data that cannot be visually discerned through image recognition.
Polytag’s UV tags embed this information directly, providing definitive rather than inferred data. This distinction becomes particularly valuable when sorting plastics with similar appearances that nonetheless require different recycling processes to maximise quality and output.
2. Centralised vs Decentralised Data Models
Greyparrot’s DeepNest operates as a centralised database containing registered packaging images and metadata. While this approach offers comprehensive coverage, it requires all participants to contribute data to a single platform. Polytag supports a more distributed ecosystem: data is encoded within the packaging itself, allowing detection units to operate independently. This model offers enhanced flexibility and respects individual data governance preferences.
3. Detection Efficiency: Fluorescence vs Image Matching
Greyparrot’s AI must compare each packaging image against an extensive library – a process that requires significant computational resources. Polytag’s approach leverages fluorescence detection, where UV tags become immediately visible under detection, allowing the data matrix to be read instantaneously. This offers advantages in high-speed recycling environments where processing efficiency is paramount.
4. Global Readiness vs Database Dependencies
Polytag’s adoption of universal GS1 standards enables deployment anywhere in the world with immediate functionality. Greyparrot’s effectiveness depends on packaging having been previously catalogued in its database. With approximately 20,000 new products launched monthly in Europe alone, maintaining comprehensive coverage presents ongoing challenges, though the company continues to expand its database capabilities.
5. Environmental Considerations: Computational Intensity
AI systems require substantial computational resources, which translates to higher energy consumption. This presents an interesting consideration: the environmental cost of the technology used to improve recycling efficiency. Polytag’s system operates without AI processing, offering a potentially lower-carbon alternative that may align more closely with broader sustainability objectives.
6. Real-Time Sortation: Embedded Data vs Lookup Systems
Polytag’s ecosystem includes sortation capability that combines near-infrared (NIR) optical sorting with UV data matrix reading.
Because packaging attributes are embedded within the matrix, sortation decisions can be made instantly – essential when conveyor belts operate at speeds up to 5 metres per second. Greyparrot’s image lookup approach introduces processing time that may present challenges in high-speed industrial environments.
The Case for Polytag: Building Tomorrow’s Recycling Infrastructure
While both technologies represent significant advances in packaging detection, the evidence suggests Polytag offers a more robust foundation for the future of recycling infrastructure. The combination of absolute data certainty, open standards compatibility, and real-time processing capabilities positions Polytag as the more scalable solution.
Perhaps most importantly, Polytag’s decentralised approach scales without expensive overheads or dependencies. This means recyclers can operate independently, new participants can join without gatekeeping, and the system remains functional even as individual components are updated or replaced.
The recycling industry needs solutions that work reliably at industrial scale, integrate seamlessly with existing infrastructure, and remain viable as the circular economy grows exponentially. Polytag’s standards-based approach, instant processing capability, and open ecosystem architecture make it uniquely positioned to meet these demands.
As we stand at the threshold of truly circular packaging systems, the technology choices we make today will determine whether we build an open, efficient, and resilient recycling infrastructure – or one constrained by computational bottlenecks and centralised dependencies. For an industry that must process millions of items daily with perfect accuracy, Polytag offers the certainty, speed, and scalability that tomorrow’s circular economy demands.