ICOSA Brewhouse

ICOSA Brewhouse designs, sells and distributes coffee accessories worldwide from Canada. It is a small team, which means everybody wears several hats and the operational work lands on people who already have another job to do.

Their stock did not live in one place. They sell through Amazon, Shopify and other channels, and every manufacturer they work with runs its own platform with its own interface. One supplier has no platform at all and sends inventory updates as photographs.

So somebody checked all of it by hand. Every channel, every manufacturer portal, and the photographs, on a daily and weekly cycle, then keyed the numbers into spreadsheets to keep a picture of what they actually held.

That had been running for years. The problem was not that nobody had noticed it. It was that nobody believed it could be automated, so it never came up as something worth fixing.

We started by mapping where inventory actually lived: each sales channel, each manufacturer platform, and the supplier whose updates arrive as images rather than data.

From there we built a pipeline that pulls each source through its API, reconciles the results, and stores them in one place. It runs on the same daily and weekly cycle the team used to work by hand, so the rhythm of the business did not have to change.

The spreadsheets stopped being the system of record and the team stopped being the integration layer. What shipped is a combination of conventional system work and AI rather than one standing in for the other, and the job left for a person is reviewing the output instead of assembling it.

Manual inventory and data work went from around ten hours a week to roughly thirty minutes of human review.

That is ten hours a week returned to a small team, spent instead on the parts of the business that need judgment and on making sure orders reach customers on time. It also changed what they think is possible: work they had written off as unavoidable turned out to be automatable.