Econex MCP lets AI assistants and agents read data, analyse and control the building – with the same per-project, per-role rights as in the SD.
Claude, ChatGPT and Copilot · Your own agent via MCP · Access control as in the SD
Claude
Energy advisor
Read energy use year to date
Read
Allowed
ChatGPT
Operations manager
Lower setpoint, AHU 360.01
Control · 21.0 → 20.5 °C
Done
Own agent
Partner, read access
Change operating mode, AHU 360.01
Control
Denied
No control right in this project
In most buildings the automation is closed to AI, or data has to be exported before anyone can analyse it. It is rare for AI to connect straight to the building through the same access control that operations use every day. That is exactly what Econex MCP does.
Claude, ChatGPT, Copilot or an agent you build yourselves.
Knows the user, the organisation and the project, and checks the rights on every call.
BMS plant, meters and sensors. Control only happens where the rights allow it.
MCP (Model Context Protocol) is an open standard that lets AI assistants connect securely to external data sources and tools. Econex MCP is one such connection: it gives ChatGPT, Claude or Copilot direct, secured read access to your buildings in Econex.
In practice it means you stop exporting CSV files and pasting numbers by hand. Instead you ask the AI in plain language, and it pulls the right data from Econex itself – energy use, temperatures, ventilation, solar production, alerts and more – and builds the analysis for you.
Rights are granted in Econex per project, just as in the SD. What the AI tool may do follows from them – and nothing else.
| Level | What the AI can do | Right in Econex |
|---|---|---|
| Read | Fetch projects and building structure, alarms, energy, comfort and ventilation, dampers, raw telemetry and calendars. | Read access per projectmcp.project.read |
| Analyse | Compare buildings, normalise for outdoor temperature and produce finished Excel, Word and PowerPoint reports. See reports made via MCP | Same read access. The analysis happens in the AI tool.mcp.project.read |
| Control | Change operating modes and setpoints, and configure plant. | Separate rights for control and configuration, granted as a deliberate choicemcp.project.operate · mcp.project.configure |
A real conversation: you ask, the AI pulls data from Econex via MCP and delivers finished files.
You don't need to know formulas or export formats. Just describe what you want.
These are actual files an AI produced by pulling data from Econex via MCP – from an energy report to a comfort analysis. Open each one to see the end result.
The preview is rendered by Microsoft Office's web viewer. You can also download the file or open it in a new tab.
Everyone connects the same way. The difference is the rights each person has in the project.
«How is the portfolio doing against this year's energy target, and which buildings are pulling it down?»
Owner and board reports with real figures, without waiting for someone to export data. Typically read access across the buildings you own.
«Which air handling units had alarms last night? Put AHU 360.01 in reduced mode until Monday.»
Ask about alarms, deviations and trends, and let the AI make changes where you hold the control right yourself. Everything goes through the same rights as in the SD.
«Give me damper status and the latest weekly test for the projects we service.»
Connect ChatGPT, Claude or Copilot to the projects you have an agreement on. Let the AI create and set up buildings, drivers, rooms and point lists – only what the building owner has opened for you.
About the partner networkEconex MCP uses OAuth. You add the address from the documentation, sign in to Econex and approve. In most AI tools there are no keys to copy or share.
An Econex administrator gives the user MCP access and chooses projects and level: read, or also control.
Add Econex MCP as a connector in Claude, ChatGPT or Copilot, or point your own agent at the same address.
The tool opens “Connect to Econex”. You sign in as usual, check the organisation and approve.
The AI now sees the tools your rights open up, and nothing more.
This page is about Econex MCP. Econex also has two AI functions used directly in the interface. All three are governed by the access centre and enabled per project.
Inside Econex
Ask questions and get help while you work in Econex. The assistant is limited to the project and your permissions.
Inside Econex, for partners
AI support for partner roles such as integrators and contractors, with its own access frame per project.
In ChatGPT and Claude
Bring the building's data into the AI tool you already use, and get analyses and reports back. The tools the AI gets follow your permissions.
We enable Econex MCP for your projects, set up the rights with you and help you get going in Claude, ChatGPT, Copilot or your own agent.