AGD Intelligence

Composed meal / grain-bowl assembly and portioning at central-kitchen stations

In Sodexo's central/commissary kitchens, prepared components (cooked proteins, grains, roasted vegetables, sauces, leafy greens, garnishes) are portioned and assembled into bowls, trays or hotel pans before QA, labeling and shelving for distribution to client sites. The task is a repeated pick-portion-place sequence per meal: select a component, portion it to a target weight or count, and lay it neatly into a vessel without crushing or smearing, often layering several items per bowl at high meal-per-hour rates. The objects are deformable and fragile with highly variable geometry (delicate leaves, soft proteins, sticky grains, fluid sauces) and inconsistent presentation tolerances. It is hard for a robot because success depends on gentle, compliant grasping and slip control rather than precise positioning, and because food variability defeats fixed-geometry handling. This is genuine line context — patents covering assembly-kitchen workflows describe exactly these manual assembly stations — but Sodexo's own stated automation investments so far target vending/cooking throughput, not dexterous assembly. 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
promising
Source
researched
Failure tol.
medium
Tactile value
high
i

What the task is

RESEARCHED · our reconstruction

In Sodexo's central/commissary kitchens, prepared components (cooked proteins, grains, roasted vegetables, sauces, leafy greens, garnishes) are portioned and assembled into bowls, trays or hotel pans before QA, labeling and shelving for distribution to client sites. The task is a repeated pick-portion-place sequence per meal: select a component, portion it to a target weight or count, and lay it neatly into a vessel without crushing or smearing, often layering several items per bowl at high meal-per-hour rates. The objects are deformable and fragile with highly variable geometry (delicate leaves, soft proteins, sticky grains, fluid sauces) and inconsistent presentation tolerances. It is hard for a robot because success depends on gentle, compliant grasping and slip control rather than precise positioning, and because food variability defeats fixed-geometry handling. This is genuine line context — patents covering assembly-kitchen workflows describe exactly these manual assembly stations — but Sodexo's own stated automation investments so far target vending/cooking throughput, not dexterous assembly.

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.