PTC ThingWorx is an industrial IoT platform for building connected applications in factories, at equipment makers and in service organisations. It models the world as Things, built from Thing Templates and Thing Shapes, each with properties, services and events. Data flows in from machines, value streams keep its history and mashups show it on screen. That flexibility has a cost: the answer to a simple question often sits behind a mashup that nobody has built yet.
Agentic CMMS lets Claude, ChatGPT and other AI agents work with ThingWorx directly. Your maintenance team asks the question, and the agent reads the Things, properties and alerts that answer it. No new mashup, no ticket to the application team.
What your AI agent can do with PTC ThingWorx
- Find Things by template. "Show me every Thing based on the Compressor template, with its site."
- Read live properties. "What are the current outlet pressure and run state of COMP-22?"
- Query property history. "How did the hourly cycle count on Filler 4 change across the last week?" The agent reads the value stream behind the property and describes the pattern.
- Review alerts. "Which Things have an unacknowledged high-priority alert right now?"
- Run a service. "Call the reset-counter service on the Bottle Inspector after this changeover." With write access, the agent runs the service and reports the result.
- Understand a model. "What properties and services does the Mixer template define?" Useful for new engineers and for anyone writing a maintenance procedure.
- Check connectivity. "Which Things at the Monterrey site are not connected, and since when?"
PTC ThingWorx with your other systems
ThingWorx tells you what a machine is doing. The people who fix it work in other systems. Bring them into the same question:
- Let a ThingWorx alert on a compressor open a work order in IBM Maximo, with the property values at the time of the alert in the description.
- For equipment you sell and service, check alerts on customer machines against open service jobs in ServiceMax or Dynamics 365 Field Service before you send a technician.
- Pull the troubleshooting page for an alert type from Confluence and add the steps to the answer, so the technician has them on the way to the machine.
For teams who built their own applications
ThingWorx models are often custom, with names that make sense mainly to the people who wrote them. The agent reads the templates, shapes and descriptions your team created, so it can follow your naming instead of asking you to learn it. The better your model, the better its answers.
Service calls deserve more care than reads, because they can change how a machine behaves. Those calls need write access, which each person grants to their own agent. Owners and admins choose which roles may use each tool, and every call is recorded in the activity log.