Manufacturing leaders have spent the past few years asking the same question: how do we get more value from AI? It’s an understandable focus. Product development teams are under pressure to deliver increasingly sophisticated products while navigating supply chain volatility, evolving regulations, compressed development cycles, and rising product complexity.

Yet many organizations are discovering something less intuitive in practice: adopting AI is the easy part. Achieving meaningful results inside engineering workflows is much harder.

The reason has less to do with the quality of today’s models than with the data they’re asked to interpret. In engineering, data without context isn’t simply incomplete. It can be misleading.

A requirement is never just a requirement. It influences a design. That design affects manufacturing decisions. Those decisions, in turn, ripple into quality, serviceability, compliance, and even future product iterations. Engineering data only makes sense through its relationships to everything around it.

Why Context Matters

When those relationships are fragmented across disconnected systems, AI can still generate answers. What it struggles to do is to confirm the validity of those answers for the specific versions, product variants and effective dates in question. It cannot fully grasp the impact of a seemingly simple change.

Consider a supplier notification that a critical electronic component is reaching end-of-life. On paper, the response looks straightforward: identify a replacement and update the bill of materials. But anyone who has worked in product development knows how quickly a single component change can create downstream impacts.

A substitute component may require updates to CAD files, simulation models, manufacturing instructions, regulatory documentation, and supplier qualification records. It may affect thermal performance assumptions, introduce new testing requirements, or even alter service procedures already in the field. What began as a simple part swap becomes a cross-functional engineering event.

AI that only sees the bill of materials might recommend a technically equivalent replacement that appears correct in isolation. But without visibility into the surrounding engineering context, it cannot account for downstream consequences. AI grounded in a connected digital thread, however, can trace those relationships, understanding not just what is changing, but what that change touches.

That distinction is where the real challenge lies. It isn’t the absence of data that limits AI in engineering environments; it’s the loss of meaning that occurs when data is disconnected from its lifecycle context.

This is why AI readiness in manufacturing is often misunderstood. Organizations tend to frame it in terms of data volume, data cleanliness, or model sophistication. Those matter. But in engineering, another requirement is easy to overlook: structure. AI must have access not only to data, but to the relationships that explain how that data connects across disciplines, systems, and time.

That is the role of the digital thread.

The Digital Thread Gives AI Context

A digital thread links requirements, design models, engineering changes, manufacturing plans, quality records, supplier data, and service information into a connected lifecycle of decisions. It preserves the relationships between that data as the product evolves.

With that structure in place, AI shifts from being a retrieval tool to something more powerful. It can interpret lineage. It can trace dependencies. It can surface implications rather than just results.

In the component example, that means going beyond identifying an alternative part. AI can help determine where the existing component is used and what a proposed change might touch. It can surface related requirements, prior engineering changes, validation activities, manufacturing dependencies, quality records, or supplier considerations that an engineer should evaluate before acting. The AI is no longer just generating a simple answer. It is helping the engineer navigate the consequences of change.
As product complexity grows, this becomes increasingly important. No individual engineer can realistically evaluate every relationship across thousands of components, requirements, changes, test results, and supplier records. AI is well suited to processing that volume, identifying patterns, and surfacing potential impacts that would otherwise be difficult to see.
But identifying an impact is not the same as making an engineering decision. Engineers still determine whether those insights make sense within the broader realities of product performance, customer commitments, regulatory obligations, and other tradeoffs.

This is where the digital thread matters. It gives AI the connected context needed to surface better insights while giving engineers the traceability needed to evaluate them. AI does not replace engineering judgment. It makes more of the relevant context available when that judgment is applied.
Even the most capable AI will fail to gain traction if engineers do not trust its outputs. In engineering environments, that trust must be earned. Teams need visibility into what data informed a recommendation, where that data came from, and whether it reflects current, governed, and reliable sources.
That traceability allows engineers to distinguish between trusted data and AI inference. Without it, AI remains a black box. With it, AI becomes something engineers can evaluate, validate, and confidently incorporate into their work.

Context Becomes the Advantage

The organizations that succeed with AI in product development will not necessarily be the ones that deploy it fastest or experiment most aggressively. They will be the ones that focus on something less visible but far more important: preserving the context that gives engineering data its meaning.

Over time, something more valuable than efficiency gains begins to emerge. The connected digital thread evolves into a form of organizational memory, capturing not just what decisions were made, but why they were made and how they influenced downstream outcomes. In that environment, AI becomes less about automation and more about continuity. It allows engineering knowledge to accumulate, connect, and persist across product generations.

As advanced AI models become more broadly available, competitive advantage will increasingly come from something competitors cannot simply buy: a trusted understanding of an organization’s own products, decisions, and engineering history.

That advantage does not require waiting for the next generation of AI. It starts with connecting engineering data today and preserving the relationships that give it meaning. The manufacturers that do this well will be in a stronger position to turn increasingly capable AI into better engineering decisions, without sacrificing the traceability and trust those decisions require.

Rob McAveney brings a lifelong passion for technology to the CTO role at Aras. For the past 20 years, he has focused that passion on building rich software platforms that solve difficult business problems for major industrial companies. Rob acts as Aras’ technology visionary and provides design oversight for future PLM technology, while remaining grounded in the realities of configuration management, systems integration, and the many other challenges of delivering enterprise software. Prior to Aras, Rob led technical sales engagements for Eigner, an early entrant in the PLM market. He began his career at Boeing, where he gained a broad understanding of engineering and manufacturing systems and processes.