Production planning laid out by AI.

Starting point

The plan gets made every day depending on department workloads.

Upholstery and sewing depend on each other: a delay in one department pushes back the rest. A planner built the plan — from ERP data, notes, conversations and their own knowledge. A missing material, an absence, or a rush order meant recalculating everything from scratch.

We started with an audit: conversations with planners, a process map, a list of input data, and a check on whether the ERP data was fit for training a model. In the next step we set the goal, the scenario and the pilot metrics.

What we built

Data flows both ways.

There was one condition: data had to flow freely both ways. The tool pulls data from the Saturn ERP — orders, deadlines, semi-finished stock, and who's on shift — and imports back the finished plan and a proposed order assignment.

How we check it

A pilot with concrete outcomes.

During the pilot we logged every case where the planner changed the system's proposal. These records show where the tool helps and where the process needed fixing.

At the end, a report is produced: how accurate the plans were, what's worth keeping permanently, and to what extent. The client confirms the pilot's results.

Tools AI / LLM tool ERP Saturn Two-way integration Rules and dictionaries EDIH hub4industry

A production plan can be tested before rollout.

We don't name the company. Numbers come from the audit and process measurement at the client.

Take the Chaos Scan Let's talk