RICHMOND, VA – Reliability remains the dominant consideration for engineers selecting components for PCB designs, while artificial intelligence has become a common part of the design workflow, according to a new survey commissioned by Weidmuller USA.

The survey of 400 North American engineers found that 79% identified high reliability and durability as the most important features for device connectivity, compared with 22% who cited price. Better performance, cost savings and improved reliability were the leading reasons respondents had tried a different product or component during the past two to three years.

Component decisions are also being made relatively early in the PCB development process. Some 43% of respondents said most component selections are finalized during either the early concept and architecture phase or schematic design, while 26% said selection typically remains iterative across multiple stages.

AI tools have meanwhile become widespread among the engineers surveyed, with 91% reporting they have used AI-based tools in their PCB design workflows. Three-quarters characterized AI primarily as a productivity and acceleration layer within existing workflows, with faster design iterations viewed as its primary current value.

Despite that adoption, respondents identified gaps between existing AI capabilities and what they would like the technology to accomplish. Write-in responses cited capabilities such as real-time error detection and fully autonomous text-to-PCB-board design as areas engineers would like to see developed further. Concerns around data security and accuracy also remain, Weidmuller said.

“This new research reinforces the importance that engineers place on improvement in performance, reliability, workflow efficiency and supply resilience, especially when these benefits are integrated directly into the PCB design process,” said John Froustet, senior director, product portfolio management at Weidmuller USA. “While AI tools are increasingly looked to for speed and accuracy, many questions around data security and accuracy of results still linger.”

The study was conducted by EETech Research during April and May 2026 through an anonymous 28-question web survey of 400 engineers in North America.