Find knowledge in your documents. Investigate opportunities for part reuse. Keep engineering decisions grounded in the evidence.
Talk through your use case
Windchill AI brings assistance into PLM work so people can investigate product information in context. Start from a real engineering question and use supported source material to help answer it.
The aim is to reduce the effort between finding information and deciding what to do with it. An answer or a reuse candidate starts an investigation: the team checks the evidence and decides the next action within its existing process.
Start from an eligible source set and a defined task.
Find relevant knowledge or investigate a potential match.
Inspect the supporting evidence before taking action.
A useful PLM experience helps people reach relevant information, spend less effort on preparation and make a clearer handoff. Use these three perspectives to assess what assistance changes in a real task.
Move from hunting across long technical documents to a focused question and a source that can be checked. The useful gain is reaching information the reviewer can actually use.
What would a useful answer let this person do next?
Explore document assistanceInvestigate repetitive preparation, such as locating potential reuse candidates, before creating another part. Assess the effort saved after the engineering comparison is included.
Does assistance reduce work after checking is included?
Explore part intelligenceUse a common review standard: identify the supporting source, the responsible person and the decision to record. AI assistance should make that handoff clearer.
Can another reviewer follow the reason for the next action?
Plan a focused pilotExplore part intelligence, document assistance and the broader agentic digital-thread direction. Each reading makes its evidence and decision boundary explicit.
AI Parts Rationalization uses configured shape indexes to identify duplicate or similar parts. Comparison tools let authorized users investigate candidates and their attributes.
Before creating another bracket, a designer can review related shapes already in the system. A promising match then needs a closer assessment of dimensions, material, function and the applicable product context.
The plugin, indexed representations, permissions and licenses must be in place. Similar shape does not establish interchangeability. Authorized users can carry a reviewed candidate into supported change or supersede workflows.
Decide whether a candidate is appropriate for reuse, then follow the governed change process.
Windchill AI Assistant supports natural-language questions and summaries over indexed text in supported documents. Source references help the reader open the material behind a response.
A team member reviewing a long engineering specification can ask a focused question, narrow the context where supported and inspect the referenced document before using the information in a decision.
The indexed content and available source modes depend on configuration and licensing. Internal source documents require the relevant Read and Download permissions. CAD geometry, image content, OCR of scanned text and general Windchill metadata are not searchable document answer content.
Review the original passage, document context and suitability of the response for the task.
PTC’s agentic digital-thread direction explores AI assistance across connected engineering, manufacturing and service work.
Consider a proposed task that crosses applications. Map the records it would read, the action it might suggest and the point where a responsible person must review the next step.
This is a broader product direction, not a list of autonomous functions included in every installation. Confirm the specific capability, supported release, deployment and approval controls with TriStar before planning a workflow.
Keep record ownership, permitted actions and human review explicit.
Windchill AI is generally available for supported environments, bringing generative AI assistance into PLM work through eligible product-document content and source-referenced responses.
For an engineering specification review, availability is the point to check whether your organization can enable the supported capability. It does not mean an existing installation already has the plugin, access or prepared information required.
TriStar can help frame the evaluation around the task you want to improve, the evidence you expect and the people who will assess it.
AI works within a wider product lifecycle. Explore the shared PLM foundation, then consider how you want that environment delivered and operated.
Bring engineering files, product structures and change records into a governed product context. Windchill gives teams a common foundation for collaboration across the product lifecycle.
Start here for the platform’s capabilities, working audiences and product-information workflows.
Explore Windchill
Bring Windchill’s PLM foundation into a SaaS delivery model. Evaluate service responsibilities, environment planning and the path from your current deployment alongside the product capabilities.
Start here for SaaS operating scope and migration planning, rather than the AI workflows described above.
AI capability, licensing and supported deployment remain separate scope decisions. Choose the product environment first, then confirm the AI requirements for that environment.
An industry context helps define the information and responsibility behind a pilot. Use these starting questions to choose source material, a review owner and the boundaries of the evaluation.
Explore a focused question in the supported specification documents available to the program team. Have the responsible engineer examine the referenced passage and its configuration context.
Which source and revision should govern the engineering decision?
Select a family of machined components and compare shape-based candidates before creating a new part. The review owner checks the functional and material differences that geometry alone cannot settle.
What would make this candidate suitable for the intended machine?
Use a permitted technical-document question to assess whether readers can inspect the supporting content. Keep the existing document-control, quality and review responsibilities in place.
Can the reviewer trace the information to the appropriate controlled source?
Clarify the source material, access and review process before expanding an AI workflow.
The Assistant works with indexed text in supported documents to help answer questions and summarize information. Parts Rationalization investigates geometric similarity between indexed parts. They address different tasks and have separate setup and entitlement requirements.
No. Its document workflow uses supported indexed text. CAD geometry, images, OCR of scanned content and general Windchill metadata are outside that searchable document content. Review the supported formats and configured sources for your deployment.
The relevant Windchill permissions remain important. A cited internal document requires Read and Download access. Test the configured workflow with the intended users and their real permissions.
It identifies a candidate to investigate. Engineering review still needs to establish whether dimensions, material, function, product context and other requirements make the part suitable.
Check the supported maintenance level, matching plugin, licenses, data preparation and deployment prerequisites. On-premises Assistant use also needs Azure resources and a paid subscription; the operating plan extends beyond the Windchill server itself.
Choose one document question or one part family, identify the permitted source material and nominate the person who will review the evidence. Agree what a useful result would look like before widening the scope.
Bring the task, the product information and the people who will make the decision. We’ll help define a practical next step.
Talk to TriStar