Warehouse Execution System
Model Context Protocol (MCP): AI Access to Your Warehouse Data
Put the warehouse data in your Warehouse Execution System to work — straight from your preferred LLM, such as ChatGPT or Claude.
AGENTIC AI: The New Partner in Intralogistics
Your people have been working with AI assistants for a long time. Just not with your warehouse data. That changes now: the identpro Warehouse Execution System additionally serves its data over MCP. ChatGPT, Claude, partner solutions or your own systems connect using an API key. And this is the part that matters: read-only.
What this makes possible:
- Answers in seconds: questions in plain language instead of clicking through screens
- Read-only access: processes, data and workflows in the WES remain untouched
- Vendor-independent: works with ChatGPT, Claude or your own LLM
- Freely combinable: warehouse data can be related to other data and analyzed directly in the chat
- No customizing: ready to use as soon as the API key has been issued
Why this matters now
The way people work has changed over the past few years: research, analysis and decision-making groundwork increasingly happen in dialogue with an assistant. What those assistants are missing is your own operational data. They know a great deal about the world — and nothing at all about your warehouse.
Opening up your warehouse data through an open standard shortens the path from question to decision. Every analysis that is a ticket to IT today becomes a conversation. And the access does not have to be rebuilt for every new tool — new models and assistants can use it right away.
Questions About Your Warehouse That Can Be Answered From Now On
Instead of screens and reports: a question in plain language, answered from your warehouse’s live data.
Presentation (table, chart, summary) is handled by the connected assistant. The Warehouse Execution System supplies the data.
See what MCP can do!
What Is the Model Context Protocol (MCP)?
The Model Context Protocol is an open, vendor-neutral standard for how AI systems access software and the data it holds. What USB is for devices, MCP is for AI and software: one standard connection instead of a one-off solution for every case. It turns the Warehouse Execution System into an open data source that AI systems can use in a standardized way.
Since 9 December 2025, the MCP standard has been maintained by the Agentic AI Foundation under the umbrella of the Linux Foundation (members include Anthropic, OpenAI, Google, Microsoft and Amazon Web Services).
What USB is for devices, MCP is for AI and software: one standard connection instead of a one-off solution for every case
Without manual effort, the LLM cannot reach the data — and may not be able to map it correctly. Every connection that is built has to be recalibrated by hand after a WES update.
One connection, and your LLM of choice has access to your warehouse data. No additional agents needed to build their own interfaces. Updates included.
One question, several queries, one explanation
Nobody has to define up front which queries a question needs. The assistant works that out itself.
Example:
Question: “Why wasn’t order 4711 completed yesterday?”
Approach: Five read-only queries through defined functions of the Warehouse Execution System.
|
STEP |
QUERY |
WHAT IS QUERIED |
|
1 |
Order |
order data and due date |
|
2 |
Material requirement |
material needed |
|
3 |
Stock |
stock level and storage location |
|
4 |
Movements |
movements in the time window |
|
5 |
Vehicles |
vehicle availability |
Answer: The three pallets required were in stock, but could not be moved between 14:32 and 15:18 — two of the three vehicles responsible were unavailable.
Result: A piece of information that today means manual research across several screens, and that no dedicated report is worth building for, because the answer is different every time.
What Data You Can Query Over MCP
Read-only queries on stock and movement data. Write functions, such as creating a transport order, are deliberately not enabled. MCP does not replace your existing interface. It adds a layer for AI-driven queries. Nothing changes about your ERP or WMS connections.
| CATEGORY | SCOPE |
| Stock | per item, storage area and load carrier |
| Material search | by item, batch or identifier |
| Storage locations | current storage location and occupancy |
| Transport orders | open and delayed orders |
| Vehicle status | availability and position of the vehicles |
| Movement history | movement data over time |
You Stay in Control
CONTROL: WHO MAY DO WHAT
- Read-only: no assistant can change data, create data or trigger movement commands.
- Access via API key: you issue and revoke the key yourself and control every connection with it.
- Attributable and switchable: every key belongs to a named connection and can be deactivated at any time.
TECHNOLOGY: WHAT THIS MEANS FOR YOUR IT
- No copy of the data: the data stays in the WES and is read at runtime. No data lake, no outdated extracts.
- Update-safe: access to defined functions instead of tables.
- On-premises operation possible: MCP access runs inside your existing environment.
Open questions about the connection?
What You Get Out of It
For partners and system providers
Integration partners connect to the Warehouse Execution System with an API key instead of specifying individual data extracts for every project. Software vendors read with ERP, TMS or MES through the same documented interface as an AI assistant: access to defined functions instead of tables, with no database access.
How to Get Started
REQUIREMENTS
- an existing installation of the identpro Warehouse Execution System
- an API key
- a connected assistant such as ChatGPT or Claude
An existing WES installation is the prerequisite. Not using it yet? Then let’s talk about both.
“Which piece of information from your warehouse do you need regularly — and how long does it take today?”
The answer to that is the use case for your first MCP pilot.
Frequently Asked Questions (FAQ)
Speak directly
