Supply chain analytics, run from ChatGPT and Claude

AI prepares your data and runs the analysis, from algorithm to dashboard. Every analysis is auditable and repeatable when the situation changes.

Get on the beta list
The question asked in Claude, with the TMS export attached: "Visualize my shipments on a SCMotif map". 1 · Asked in Claude
The European network SCMotif built from the export: six producers feeding two hubs in Rotterdam and Vienna, ten distribution centres from Lisbon to Helsinki. 2 · Built by SCMotif

SCMotif plugs into the agent you already use, over the Model Context Protocol.

ChatGPT · Claude
you ask here
SCMotif
MCP server, runs the analysis
Workspace
workflows, scenarios, dashboards

From your question to a result you can defend

One run on four raw SAP exports.

The question 1 / 3

You ask, the agent builds, and asks back when a decision is yours

Four raw SAP tables and one sentence. Three modelling decisions came back as questions before anything was built.

The chat: the question "Build a network from these SAP exports and work out the cost to serve each customer", and the answer naming the four SAP tables and what is wrong with them before any modelling decision is taken.
The workflow SCMotif built: four SAP source files feeding a cleaning step, five mapping branches for locations, flows, lanes and demand, then the scenario and the freight allocation.
A dashboard headed "Cost to serve by customer, SAP Jan to Feb 2024": total cost to serve €16,438 allocated to 20 customers from 27 billing lines, revenue €177,521, freight at 9.26% of revenue, net contribution €161,083, and 40,052 kg shipped with a note that three lanes have no stated material weight and are excluded, over two bar charts ranking customers by freight share of revenue and by cost to serve.
MCP server opens with the beta

Connect your agent

SCMotif runs as an MCP server. Your agent connects to it and gets tools it doesn't have on its own.

Flow maps

See where the volume actually goes.

Center of gravity

Where a site should sit, given the flows.

XYZ segmentation

Sorts items by how steady their demand is, before anything is planned around them.

What-if scenarios

Build several and compare them side by side.

Network optimization

Solves a mixed-integer model instead of guessing the answer.

From the blog
Outside-in teardown August 2026

When does discount retailer Action need a Romanian warehouse?

We rebuilt the local DC decision from public data: break-even at ~56 stores (55–70 across tested assumptions), and a best site that flips as the network grows.

Read the teardown →
SCMotif scenario map: a Ploiești DC serving 150 Romanian stores

We're opening a private beta

For people who deliver quantitative network work and want their agent to do more of it.

  • Access to the hosted workspace and the MCP server, free while the beta runs. It becomes a tiered subscription after that.
  • You bring one real analysis and the data it actually needs. Not a demo dataset.
  • We keep your data in our EU workspace and never pass it to a model provider. What your own agent sends to its provider stays under your agreement, not ours.
  • If it works, we'd like to name you. If it doesn't, we want to hear why.
Get on the beta list contact@scmotif.com

SCMotif is built by Lovis Anderson, previously product owner for network design software at OPTANO, a Kearney company. Who builds this →