In Irish salmon processing plants, the heavy lifting starts before anything else. Pallets of fish arrive from Norway, Iceland and the Faroe Islands stacked nine or ten layers high, each polystyrene box holding up to 30 kilos of salmon and ice. Two people spend the shift pulling the boxes off the pallet by hand, cutting the straps, lifting the lids and tipping the fish onto the line. It’s heavy, cold, repetitive work that gets harder to staff every year: it's not work Irish staff want, so plants lean on people from abroad, who don't always stay – and wages keep climbing either way.

It's the same job in nine or ten plants around Ireland. So when Bord Iascaigh Mhara (BIM), the country's seafood development agency, went looking for a way to automate it, it wasn't solving one plant's problem – it was testing a route the whole sector could follow.

A better way to find the options

Most processors have little in-house robotics experience. What can realistically be automated, how to specify it, and which suppliers can deliver isn't obvious – and there are no robots in Irish seafood processing today to copy. That gap, between wanting to automate and knowing how to start, is where most projects stall.

BIM's role is to develop the Irish seafood industry, from catch to market. It doesn't buy equipment for processors; it finds and vets options and leaves the commercial decision to the business. "Our job is to give them options," says Tomás Cooper, a business development executive at BIM. But finding good ones is hard: canvassing suppliers one by one is slow, and a small team can only know so much of a global market. So BIM turned to HowToRobot – an independent platform that lets a company scope an automation project and source suppliers from a global network, then compare what comes back. "We found suppliers through this platform that we wouldn't have found otherwise," Cooper says.

An expert eye turns a messy process into a clear brief

An automation engineer from HowToRobot's network visited two of the plants and watched the process closely – how the pallets came out of the cold store, how the fish reached the hopper, and the details that make this harder to automate than it looks: the salmon must be handled gently to avoid bruising, and the flimsy polystyrene boxes, with their straps and lids, are awkward to grip and separate. From that he produced a supplier-ready specification – the kind of document BIM would have struggled to write itself. "Specifying that out would have been difficult for us," Cooper says.

Four suppliers, four different ways to solve it

Published anonymously to the network, the brief drew four suppliers – and four genuinely different answers to a problem BIM had assumed had one. One proposed two robots working together, guided by machine vision, to de-strap and empty the boxes. Another offered a single high-speed robot built for washdown hygiene, handling two boxes at a time. A third designed a compact cell with a handling table, built to tip the fish gently and fit a tight footprint. All four automated the full sequence – cutting straps, removing lids, emptying the fish, stacking the empty boxes – and all quoted delivery within a year.

The prices varied widely – a spread of more than 150% between lowest and highest – a range no processor would ever see from a single supplier, and the clearest argument for comparing several. "The suppliers coming up with those different options – that's the best part of it," Cooper says.