The systems are positioned less as generic GenAI assistants and more as domain-specific, workflow-embedded “agents” that operate within engineering, PLM, simulation, and shop floor processes, where they take on specific work steps. This not only changes the productivity of individual roles, but also the controllability of development and industrialization programs, because knowledge, rules, and context are finally being cast into repeatable processes.

New category of AI-supported experts

A very recent signal comes from Dassault Systèmes: At the French company's 3DEXPERIENCE World 2026, “Virtual Companions” were introduced as a new category of AI-supported experts in the 3DEXPERIENCE world. According to the French company, “Aura” is already available, while ‘Leo’ and “Marie” are to follow in the course of the year. The assistants are not only intended to provide answers, but also to support users in the context of industrial know-how when creating, testing, and validating innovations – in other words, precisely where engineering teams today invest time in search work, variant comparison, and methodological steps. The three agents are designed for different areas of expertise—such as engineering, simulation, or materials science—and deliberately provide different perspectives appropriate to their roles in response to the same question, as a design engineer, a calculation engineer, and a materials expert would do.

AI function that intervenes directly in parts management workflows

At the same time, the focus is shifting toward “co-pilot in PLM” because that is where the economic leverage is particularly great: duplicates, inconsistent master data, and slow parts searches are classic obstacles to reuse, variant management, and time-to-market. On January 27, 2026, PTC announced “Windchill AI Parts Rationalization” – an AI function that intervenes directly in parts management workflows, identifies similar/duplicate parts, and supports consolidation via existing change processes. The strategic value for decision-makers lies in scalability: when part findability, duplicate avoidance, and data quality are systematically improved, rework rates and follow-up costs in purchasing, warehousing, and manufacturing decrease – without each department maintaining its own “Excel truth.”

New solution for change management in engineering

It is also noteworthy that copilots are increasingly being presented as “agents” for specific process steps – not just as text assistants. With a view to speeding up market launch, PTC announced a collaboration with Lamborghini S.p.A. a few weeks ago to improve the entire product lifecycle through a centralized database that will combine design, construction, and operation with advanced AI capabilities. In addition, as part of its collaboration with Microsoft, PTC is introducing a new solution for change management in engineering – an AI-powered agent for optimizing the design workflow, which is also expected to offer a wide range of applications beyond the automotive sector. This addresses a real bottleneck for many companies: change management ties up highly skilled capacity because the impact on bills of materials, manufacturability, service, and compliance is often compiled manually. A copilot that makes dependencies transparent from the data foundation and suggests next steps can shorten throughput times without undermining governance – provided that role rights, traceability, and approvals remain strict.

From specialist bottleneck to standard tool for variant decisions

A third, very dynamic strand is copilots for simulation and calculation. On February 3, 2026, SimScale and Hexagon announced that they would combine Marc Solver—a finite element analysis (FEA) solver for demanding structural simulations—with a new generation of “agentic AI” to make sophisticated nonlinear simulation more widely available in the cloud. The point for the industry is clear: as setup, meshing, and iteration work becomes more automated, high-quality simulation moves earlier into the development process – and becomes less of a specialist bottleneck and more of a standard tool for variant decisions.

Transforming implicit expert knowledge into repeatable, auditable work patterns

Ultimately, Copilot is establishing itself as a “skill multiplier” in manufacturing-related engineering domains, especially where skills shortages and process complexity converge. In a recent CIMdata commentary, Hexagon Copilot is described as a building block that closes knowledge gaps by providing simplified, context-sensitive help and best practice guidance within applications and feeding back learnings from the field in the form of guidelines. For decision-makers, this is more than just training: it is a way to translate implicit expert knowledge into repeatable, auditable work patterns—a core problem when trying to scale processes across multiple locations in a stable manner.

Nothing good happens unless you do it!

If you want to summarize the latest developments in industrial copilots in a nutshell, a quote from German writer Erich Kästner helps: “Nothing good happens unless you do it!” In other words, by 2026, industrial copilots will create value where they not only talk the talk, but also walk the walk—with domain context, clean data connectivity, and clear communication. you do it!" In other words, in 2026, Industrial Copilots will create value where they not only talk the talk, but also reliably get the job done – with domain context, clean data connectivity, and clear accountability. Those who invest now should decide based on three strict criteria: Is the copilot deeply embedded in the respective workflow (PLM/engineering/OT)? Are its recommendations comprehensible and version-secure? And can it be operated across locations without creating shadow IT? Those who can answer yes to these three questions will not only succumb to the fascination of the pilot, but will also benefit tangibly from scalable productivity.