Industries

Supply Chain & Logistics

Tracking, routing and warehouse tooling that removes the phone call.

Every logistics operation has a spreadsheet and a WhatsApp group holding it together. The goal is not to digitise the whole chain at once — it is to make one leg of it observable end to end, then let the savings fund the next. We start where the money leaks: empty miles, partial loads, and inventory nobody can see.

Google Cloud for supply chain

We build the chain on BigQuery, Pub/Sub and Vertex AI forecasting, with SAP and Oracle kept as the system of record rather than replaced. The pattern is the same everywhere: connect operations data first, then apply AI to the exceptions.

  1. 01 · Sense

    Shipment events · Pub/Sub

    Demand and movement read on the same clock.

  2. 02 · Predict

    Vertex AI forecasting

    A forecast a planner actually uses, not one that gets filed.

  3. 03 · Move

    Route and load building

    A plan the dock can still execute at 6am.

  4. 04 · Prove

    BigQuery

    Scope 1 and load efficiency from the same trip records.

46.8%
potential reduction in human-serviced contacts from gen AI automation across operations
Source: Industry benchmark, paraphrased

Where we help

Visible inventory, one truth
Stock positions across sites, 3PLs and contract manufacturers reconciled into a single read model, with the exceptions surfaced rather than averaged away.
Routing and load building
Route and load planning integrated with what the warehouse actually picked. Most "optimisation" fails because the plan and the dock disagree by 6am.
Exception-first operations
Alerts on the shipments that will miss, not a wall of green tiles. Demand sensing and forecasting where the forecast is used by a human, not filed.

Working on something in Supply Chain & Logistics?

A 30-minute call is usually enough to work out whether this is a quick win or a rebuild — and we will tell you if it is the second one before you spend anything.