AI waste intelligence for recycling plants and packaging brands.
Photo credit: Greyparrot
Greyparrot is a London AI company that measures what actually flows through materials recovery facilities. Analyzer cameras sit above conveyor belts, classify plastics, fibre, metals, glass and other items in real time, and feed a cloud portal used by operations, engineering, quality and commercial teams. Sync pipes that live data into existing plant systems; Deepnest, launched in 2025, shows consumer-packaged-goods brands how their packaging sorts in working plants rather than in lab models.
The company was founded in 2019 by Mikela Druckman (CEO), Ambarish Mitra and Nikola Sivacki, who previously scaled computer vision at Blippar. The name nods to the African grey parrot’s mimicry: the models learn to recognise the objects people throw away. Customers cited on the company site include Veolia, Suez, Biffa, WM, Bollegraaf and brand trials with Unilever, Amcor, Asahi, L’Oréal and Kenvue.
Greyparrot’s model is measurement as infrastructure. Most recovery facilities still audit a tiny fraction of their stream by hand. Analyzer units replace that sample with continuous, item-level composition data so plants can cut residue, defend bale quality and report to regulators with evidence. The same detections become a feedback loop for brands and packaging producers: if a format is invisible or unsortable in the field, design and EPR strategy can change before more units are placed on shelves.
The company sells software and analytics rather than competing as a robotic picker. Strategic hardware partners such as Bollegraaf embed the vision layer in new plants, while Deepnest sells insight to the upstream companies that decide polymer, label and format choices.
* Founded 2019 in London; ~50 people from 23 nationalities (company site)
* $60M total funding reported after a $27M Series B in July 2026 led by Omar Mir (prior: $2.2M in 2020, $11M Series A in 2022, $12.8M Bollegraaf strategic round in 2024)
* 250+ Analyzer units across 65+ facilities in 20+ countries
* 111+ material classes in the recognition library
* 52 billion waste objects analysed and 477 billion detections in 2025; one trillion detections claimed by July 2026
* TIME Best Inventions 2025; ISO 27001 and Cyber Essentials
* Documented plant outcomes on greyparrot.ai include a purity drop caught in 12 minutes (cited £47k reprocessing saving) and a residue cut from 30% to 20% at KSI Recycling
Continuous belt analytics
Analyzer units classify items on the line instead of relying on sub-0.1% manual samples.
Plant operations portal
Inspection, alerts and reporting dashboards help MRFs react when composition or purity shifts.
Deepnest for brands
Packaging teams see how SKUs sort in real recovery systems, including PPWR/EPR-style benchmarks.
Integrator partnerships
A 2024 Bollegraaf/Van Dyk partnership embeds Greyparrot in newly built recycling plants.
Security posture
The company states ISO 27001 and UK Cyber Essentials coverage for waste-intelligence data.
Focus areas
Locations
