Looking for Insights on New AI Features in ThingWorx 10

[leekhil_kataray]

Hi Community,

We recently upgraded our platform to ThingWorx 10, which includes new AI capabilities, and I’m looking to better understand how these features can be utilized in real industrial applications.

For additional context, we are using ThingWorx 10 together with SCPA 164, so I’m also curious about how these AI features align with or enhance SCPA-based solutions.

Specifically, I’m interested in learning:

• How the new AI/ML features in ThingWorx 10 can help improve or optimize existing applications

• What use cases these AI features are intended for, and where they provide the most value

• How other developers are integrating or embedding AI into mashups, workflows, or analytics

• Best practices, architecture guidance, or known limitations when using AI features in ThingWorx 10

• Whether any extra configuration or supporting components are required, especially when working with SCPA 164

If anyone has hands-on experience with the AI features in ThingWorx 10 or has already implemented them in production environments, I would really appreciate your insights, examples, or recommendations.

Thanks in advance!

[Rocko]

Can you specify which features you refer to when you say "the new AI/ML features in ThingWorx 10".

The newest release of TWX is 10.0 and this release does not contain added AI features to my knowledge.

There is ThingWorx Analytics but it was available in previous releases as well.

You can use a LLM with ThingWorx, check here: https://community.ptc.com/t5/ThingWorx-Developers/Using-an-LLM-with-ThingWorx/td-p/1019777

and here: https://github.com/thingworx-field-work/Thingworx-LLM/releases/tag/V2.0

But this does not necessarily need ThingWorx 10.0.

10.1, to be released in December, will bring some new AI features, and there is also a partnership with Leela you can read about here or on LinkedIn. Your PTC sales rep and TWX product mgmt would be the right ones to talk to about this.

[leekhil_kataray]

The feature I was referring to was demonstrated during the ThingWorx 10 webinar, where a user could ask an OEE-related question—something like, “Provide me the list of equipment with OEE > 75.” I’m not exactly sure of the exact question asked in the demo, but it was along those lines. 

[Rocko]

You can review the webinar here:

https://www.ptc.com/en/events/thingworx-10-launch-webinar

The section was called “ThingWorx Vision” and the AI Assistant was a private preview, but I think they make the use cases clear.

[leekhil_kataray]

Thanks for the webinar — very useful!

I was specifically referring to the AI Assistance section that starts at around 35:00 in the video (the part showing a user asking “list all available machines” and “compare performance of the top five machines” with charts). I checked the ThingWorx 10.0 launch replay page but didn’t find that specific chat/assistant demo called out there.

Could you confirm whether that AI assistant demo is part of the ThingWorx 10.0 release or if it belongs to the 10.1 private preview (which mentions embedded chat/AI agents)? If it’s in the 10.1 preview, how can I get access or sign up for the private preview and any required licensing?

Also: what are the prerequisites on my ThingWorx environment to support the demoed use cases — e.g. required ThingWorx version, LLM integration approach (bring-your-own LLM or PTC-managed), and any data/modeling requirements for “list machines / compare top 5 / chart” scenarios?

[Rocko · Accepted solution]

"Private preview" would need a discussion with ThingWorx Product Management. Since 10.1 is around the corner, I can't say if they still accept applications. I have no details on the requirements, but there is a posting about the private preview here:

https://community.ptc.com/t5/IoT-Connectivity-Tips/ThingWorx-10-1-Private-Preview-now-available/tac-p/1044801#M2251

[leekhil_kataray]

Thanks, Rocko.
I’ve added my comment to the post above regarding the AI Assistance feature in 10.1.