Real cases

Industrial AI for high-volume breweries

An industrial brewery is two factories coupled together. Upstream, a continuous process: brewhouse, fermentation, maturation, filtration, tank cellar. Downstream, a discrete, multi-reference process running at extreme speed: several packaging lines alternating brands, formats, returnable glass, new glass, can and keg. And around it all, massive logistics with a returnable packaging pool constantly entering and leaving the market.

The pain points of this sub-sector repeat from plant to plant: line startup is still documented on paper; process machines fifteen or twenty-five years old hold vital data trapped in panels with no network; lab results live in spreadsheets that somebody retypes into the ERP weeks later; the yard burns carrier hours on keying; batch and expiry coding is still typed by hand on a panel; and every brand changeover is signed off without objective evidence.

This page holds the 12 canonical IRIS cases applied to this sub-sector, grouped by piece: Connect (capturing what is lost today), Edge (AI vision on the packaging line) and Agents (plant memory turned into decisions). Each case explains the concrete problem, the iLEAN piece that solves it, the measurable before/after and a conservative payback range flagged as an estimate to validate.