YieldMind

YieldMind: Better e-waste sorting starts with the line you already have.

YieldMind develops AI-powered sorting hardware designed to fit existing electronics recycling lines. Our retrofit helps operators identify devices and route them for recovery.

Book a site visit How it works

THE RETROFIT

Add sorting intelligence to your existing conveyor.

YieldMind combines camera-based device identification with hardware designed around your facility’s layout and workflow. The system is designed to recognise selected device categories as they move along the belt and help route them based on how your facility handles each category.

Built around real operators’ lines, and now being piloted at a Swiss recycling centre.

The floor today runs on eyes and hands.

Four problems · Observed on site · Each one answered below
01

Mixed waste, sorted by hand

Pulling genuine e-waste out of incoming tonnage is done item by item. The facility moves at human speed.

What it costs The whole line runs at the speed of the slowest pair of hands.
02

One battery, one week down

On the floors we visited, devices are opened by hand to find cells. One that reaches the shredder means a day to a week offline, and a fire risk.

What it costs 1 day–1 week Offline when a single battery reaches the shredder.
03

An audit that never runs

Swiss rules call for a record of the stream and what's worth disassembling. On the floors we visited, it lives in people's heads.

What it costs 2–3× / year Metal extraction runs, so sorting mistakes stay hidden for months.
04

Circuit boards, sold blind

Gold- and palladium-rich boards get shredded with low-grade ones. The premium is lost at smelter settlement.

What it costs CHF 4–6/kg Price gap between high- and low-value streams, lost when boards are sold mixed.

SOURCE · Site visits and operator interviews at Swiss WEEE facilities, 2026. Figures as reported by operators.

How it works

Not just a dashboard. It sees, decides and moves the item.

  1. 01 / SEE

    Camera over the belt

    Mounted on your existing conveyor, it captures every item as it passes, dirt, dents and all.

  2. 02 / DECIDE

    Real-time classification

    A vision model trained on real intake-condition waste decides what each item is before it leaves the frame.

  3. 03 / MOVE

    Mechanical diversion

    The decision fires a diverter that physically moves the item to the right bin or belt. Diverter arms are fitted to what your floor sorts, and can be changed.

  4. 04 / LEARN

    Logged and improving

    Every decision is recorded for audit. Every operator correction becomes training data.

Any one of these pays for it.

Four value pillars · One for each problem
Fixes Problem 01 · Mixed waste, sorted by hand 01 · THROUGHPUT

The belt stops moving at human speed. Operators move to work that needs judgement.

Fixes Problem 02 · One battery, one week down 02 · SAFETY

Batteries flagged before the shredder. One prevented stop can cover the system.

Fixes Problem 03 · An audit that never runs 03 · AUDIT

The record the rules ask for becomes a byproduct of running the line.

Fixes Problem 04 · Circuit boards, sold blind 04 · VALUE

High-value devices and boards kept apart, so the smelter pays for what's really there.

Run a WEEE line? Let's walk your floor.

We usually start with an hour on your floor, watching how intake and sorting run today.

Book a site visit