Artificial intelligence is becoming an important part of Product Lifecycle Management (PLM), but for engineering and manufacturing companies, AI adoption raises a fundamental question: Who controls the data?
When discussing AI with engineering and manufacturing organizations, we often find that their first question is not necessarily what AI can do for them.
Instead, it is something more fundamental:
"What happens to our data?"
This is a legitimate concern, particularly for companies whose competitive advantage depends on proprietary product designs, manufacturing knowledge, and intellectual property.
AI offers significant opportunities to improve engineering productivity, accelerate decision-making, and make better use of product information. However, these benefits should not come at the expense of data security, intellectual property protection, or organizational control.
At Nora IPLM, we believe companies should be able to adopt AI on their own terms, within their existing security and governance requirements.
Why Data Governance Matters in AI-Powered PLM
PLM platforms manage some of the most valuable and sensitive information within an organization.
This includes:
- Product designs and CAD data: Proprietary geometries, technical drawings, and engineering specifications.
- Bills of Materials (BOMs): Product structures, component relationships, and configuration information.
- Requirements and specifications: Technical requirements, system definitions, and compliance documentation.
- Manufacturing processes: Production instructions, manufacturing methods, and process knowledge.
- Supplier information: Vendor relationships, sourcing information, and component specifications.
- Quality and change records: Engineering changes, nonconformances, corrective actions, and quality documentation.
- Engineering knowledge: Design decisions, technical discussions, lessons learned, and historical product information.
Individually, these records contain sensitive information. When connected through a PLM platform, they form an even more valuable source of engineering knowledge.
This interconnected information is also what makes AI particularly promising in PLM.
An AI system that understands product structures, engineering relationships, change history, and lifecycle context could help teams locate information faster, investigate dependencies, understand the impact of changes, and make more informed decisions.
However, as AI gains access to richer engineering context, the importance of controlling how that information is processed and where it is stored increases significantly.
The more valuable the engineering context becomes, the more important data governance becomes.
The Challenge of AI in Regulated and Security-Sensitive Industries
Not every organization can send engineering data to an external AI service.
In industries such as aerospace, defense, automotive, electronics, and industrial manufacturing, organizations may operate under strict contractual, regulatory, or internal security requirements.
Certain programs may prohibit transferring technical information outside approved environments. Others may require systems to operate within dedicated networks, private infrastructure, or completely isolated environments.
In some cases, internet access is restricted or unavailable by design.
These constraints create an important challenge for AI adoption.
Even when an AI capability provides clear operational benefits, an organization may be unable to use it if doing so requires transferring sensitive product information to external infrastructure.
A PLM platform should not force companies to choose between accessing AI capabilities and complying with their security policies.
Instead, AI should be optional, controllable, and compatible with the organization's infrastructure and data governance model.
Our Approach: Giving Organizations Control Over AI in Nora IPLM
At Nora IPLM, we recognize that organizations have different requirements for AI adoption.
Some companies are comfortable using externally hosted AI services. Others prefer private infrastructure. Some cannot permit AI functionality at all within particular environments.
To address these differences, Nora IPLM supports two important capabilities.
1. A Master AI Switch for Complete Control
Nora IPLM includes a master AI switch that allows organizations to disable AI activity across the platform.
This is more than simply hiding an AI interface or disabling individual AI features.
When AI is not permitted, organizations can turn it off completely.
This provides a clear governance mechanism for companies operating under strict security policies or managing environments where AI usage has not been authorized.
It also allows organizations to introduce AI according to their own adoption strategies, rather than making AI a mandatory part of their PLM deployment.
AI should be a capability organizations choose to enable, not a requirement imposed by their software provider.
2. Support for Company-Hosted LLM Infrastructure
The second capability addresses organizations that want to benefit from AI while retaining control over their computing infrastructure and engineering information.
Nora IPLM can be integrated with Large Language Model (LLM) infrastructure hosted and controlled by the customer.
This provides a path toward AI deployments where model execution and engineering data processing take place within customer-managed environments.
Depending on the infrastructure and deployment configuration, this can support:
- Private AI environments: AI infrastructure managed within an organization's approved security boundaries.
- On-premise AI deployments: LLMs hosted on infrastructure owned or controlled by the customer.
- Offline AI environments: Locally hosted models operating without external AI service connectivity.
- Organization-controlled model selection: The ability to use an LLM environment that aligns with internal IT, security, and governance policies.
For organizations with strict intellectual property protection requirements, these deployment options can be particularly important.
They create opportunities to use AI with sensitive engineering data without making external AI processing a mandatory dependency.
Why This Matters for the Future of AI in PLM
The long-term potential of AI in PLM extends well beyond basic chat interfaces or document summarization.
As AI becomes more deeply connected to engineering information, it could help organizations understand complex relationships across products, BOMs, requirements, workflows, documents, and change processes.
Consider an AI assistant that can identify which components might be affected by an engineering change, retrieve relevant historical decisions, or explain how a requirement relates to a product configuration.
These capabilities become more valuable when AI can work with connected engineering information rather than isolated files.
This is where the digital thread becomes especially relevant.
By connecting product data, processes, and engineering decisions throughout the lifecycle, PLM provides the context AI needs to deliver more meaningful results.
But that also means AI may need access to some of the organization's most sensitive information.
As the relationship between AI and the digital thread becomes stronger, security and governance must remain fundamental parts of the architecture.
Companies need control over which AI systems can access their engineering information, where those systems operate, and whether AI is enabled in the first place.
AI Adoption Should Happen on the Customer's Terms
There is no single AI deployment model that works for every engineering and manufacturing organization.
A smaller manufacturer may prefer the convenience of hosted AI services. A large industrial company may operate its own AI infrastructure. A defense contractor may require AI systems to function entirely within a restricted network.
Each approach comes with different technical, operational, and security considerations.
Enterprise PLM should accommodate these differences rather than prescribe a single model.
At Nora IPLM, our approach is based on a simple principle:
Organizations should remain in control of where their data goes, which AI environment is used, and whether AI is permitted at all.
AI has tremendous potential to change how companies manage product information, engineering knowledge, and lifecycle processes.
However, meaningful enterprise adoption requires more than advanced AI capabilities. It also requires trust, transparency, and control.
Companies should not have to choose between AI innovation and data governance.
Enterprise PLM should support both.
Explore Nora IPLM
Discover how Nora IPLM connects engineering data, product structures, workflows, and lifecycle processes while supporting different deployment and governance requirements.
