AGD Intelligence

Dessert garnishing and decoration on soft surfaces

Bakkavor produces chilled desserts that require finishing touches — placing fruit, chocolate pieces, cream rosettes or decorative toppings onto soft, fragile dessert surfaces (mousse, cheesecake, sponge). The decorative items are small, delicate and variable, and the substrate is easily dented or cracked if the placing force is wrong. The task is a finishing step before lidding/packing and is typically done by hand to achieve premium, brand-consistent presentation for retailers like M&S and Waitrose. It is hard to automate because the robot must place lightweight, fragile garnishes precisely onto a compliant surface, controlling contact so it neither drops the garnish nor sinks it into or damages the dessert. High product variety and frequent seasonal SKU changes raise the dexterity bar further. We identified this through our own research; we have not confirmed the specifics with the customer directly. This page is our researched read — a starting point for that conversation.

Readiness
build now
Demand
weak
Source
researched
Failure tol.
medium
Tactile value
very high
i

What the task is

RESEARCHED · our reconstruction

Bakkavor produces chilled desserts that require finishing touches — placing fruit, chocolate pieces, cream rosettes or decorative toppings onto soft, fragile dessert surfaces (mousse, cheesecake, sponge). The decorative items are small, delicate and variable, and the substrate is easily dented or cracked if the placing force is wrong. The task is a finishing step before lidding/packing and is typically done by hand to achieve premium, brand-consistent presentation for retailers like M&S and Waitrose. It is hard to automate because the robot must place lightweight, fragile garnishes precisely onto a compliant surface, controlling contact so it neither drops the garnish nor sinks it into or damages the dessert. High product variety and frequent seasonal SKU changes raise the dexterity bar further.

To confirm with the customer

Is this the actual task and sequence? What are the real tolerances, cycle rate, and reject criteria, and which steps are today's manual bottleneck? Answering these is what turns this from a researched signal into a validated use case.